iPPBS-Opt: A Sequence-Based Ensemble Classifier for Identifying Protein-Protein Binding Sites by Optimizing Imbalanced Training Datasets

Knowledge of protein-protein interactions and their binding sites is indispensable for in-depth understanding of the networks in living cells. With the avalanche of protein sequences generated in the postgenomic age, it is critical to develop computational methods for identifying in a timely fashion the protein-protein binding sites (PPBSs) based on the sequence information alone because the information obtained by this way can be used for both biomedical research and drug development. To address such a challenge, we have proposed a new predictor, called iPPBS-Opt, in which we have used: (1) the K-Nearest Neighbors Cleaning (KNNC) and Inserting Hypothetical Training Samples (IHTS) treatments to optimize the training dataset; (2) the ensemble voting approach to select the most relevant features; and (3) the stationary wavelet transform to formulate the statistical samples. Cross-validation tests by targeting the experiment-confirmed results have demonstrated that the new predictor is very promising, implying that the aforementioned practices are indeed very effective. Particularly, the approach of using the wavelets to express protein/peptide sequences might be the key in grasping the problem’s essence, fully consistent with the findings that many important biological functions of proteins can be elucidated with their low-frequency internal motions. To maximize the convenience of most experimental scientists, we have provided a step-by-step guide on how to use the predictor’s web server (http://www.jci-bioinfo.cn/iPPBS-Opt) to get the desired results without the need to go through the complicated mathematical equations involved.


Introduction
Individual proteins rarely function alone. Most proteins whose functions are essential to life are associated with protein-protein interactions [1]. Actually, these kinds of interactions affect the biological processes in a living cell. To really understand protein-protein interactions, however, it is indispensable to acquire the information of protein-protein binding site (PPBS). Despite many studies on the binding site of a protein or DNA with its ligand (small molecule) have been made [2][3][4][5][6][7][8], relatively much less studies have been conducted on PPBS, particularly based on the sequence information alone. It is both time-consuming and expensive to determine PPBS purely based on biochemical experiments. Facing the enormous number of protein sequences generated in the postgenomic era, it is highly desired to develop computational methods to identify PPBSs for uncharacterized proteins so that they can be timely used for both basic research and drug development, such as conducting mutagenesis studies [9] and prioritize drug targets.
Given a protein sequence, how can one identify which of its constituent amino acid residues are located in the binding sites? Actually, considerable efforts were made to address this problem [10,11]. Although the aforementioned works each have their own merits and did play a role in stimulating the development of this area, further work is needed due to the following shortcomings: (1) The datasets used by these authors to train their prediction methods were highly imbalanced or with a strong bias; i.e., the number of non-PPBS samples was significantly larger than that of PPBS samples; (2) None of their prediction methods has a publicly accessible web server, and hence their practical application value is quite limited, particularly for the majority of experimental scientists.
The present study is initiated in an attempt to develop a new PPBS predictor by addressing the aforementioned shortcomings. According to the Chou's 5-step rule [12] and the demonstrations in a series of recent publications [13][14][15][16][17][18][19][20], to establish a really useful sequence-based statistical predictor for a biological system, we should make the following five aspects crystal clear: (1) how to construct or select a valid benchmark dataset to train and test the predictor; (2) how to formulate the biological sequence samples with an effective mathematical expression that can truly reflect their intrinsic correlation with the target to be predicted; (3) how to introduce or develop a powerful algorithm (or engine) to operate the prediction; (4) how to properly perform cross-validation tests to objectively evaluate its anticipated accuracy; (5) how to establish a user-friendly web-server that is accessible to the public. Below, we are to address the five procedures one-by-one.

Benchmark Dataset
Two benchmark datasets were used for the current study. One is the "surface-residue" dataset and the other is "all-residue" dataset, as described below. The protein-protein interfaces are usually formed by those residues, which are exposed to the solvent after the two counterparts are separated from each other [21]. Given a protein sample with L residues as expressed by: P " R 1 R 2 R 3 R 4 R 5 R 6 R 7¨¨¨RL (1) where R 1 represents the 1st amino acid residue of the protein P, R 2 the 2nd residue, and so forth. The residue R i pi " 1, 2,¨¨¨, Lq is deemed as a surface residue if it satisfies the following condition: where ASA(R i |P) is the accessible surface area of R i when it is a part of protein P, ASA(R i ) is the accessible surface area of the free R i that is actually its maximal accessible surface area as given in Table 1 [22], and φ pR i q is the ratio of the two. Furthermore, the surface residue R i is deemed as interfacial residue [23] if: ASA pR i |Pq´ASA pR i |PPq ą 1Å 2 (3) where ASA(R i |PP) is the accessible surface area of R i when it is a part of protein-protein complex. For a given protein, we can use DSSP program [24] to find out all its surface residues based on Equation (2), and use PSAIA program [25] to find all its interfacial residues based on Equation (3).
If only considering the surface residues as done in [26] for the 99 polypeptide chains extracted by Deng et al. [10] from the 54 heterocomplexes in the Protein Data Bank, we have obtained the results that can be formulated as follows: Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed. In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Due to the two layers structure, the algorithm computational complexity becomes the prime concern. For the simplicity and high processing speed, saliency detection methods in the frequency domain are surf " Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was propo such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fou transform(HFT) [7]. With regards to small target detection, frequency domain method is quite diffe from other methods. It transforms the airspace information to the frequency domain, defines signifi target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on parameters. It calculates the difference between the original signal and a smooth one in the log ampli spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach det small targets from the reconstruction that is calculated only by the phase spectrum of the input sig It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational c HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spe filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstan e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, riv bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the hu visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing sys consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The re this paper is organized as follows. In Section 2, we describe the framework of the proposed algori for small target detection. In Section 3, we present the experimental results. Section 4 is the conclu of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual atten selection mechanism to make it easy to understand and analyze. On the other hand, as a compo of low-level artificial vision processing, it facilitates subsequent procedures by reducing computati cost, which is a key consideration in real-time applications. Based on the above knowledge, we prop a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection st a saliency map(SM) is obtained and the most salient region is picked up to improve detection sp In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Due to the two layers structure, the algorithm computational complexity becomes the prime conc For the simplicity and high processing speed, saliency detection methods in the frequency domain s urf ď Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform wa such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomp transform(HFT) [7]. With regards to small target detection, frequency domain method is qui from other methods. It transforms the airspace information to the frequency domain, defines target and tests in the frequency domain. While, spectral residual (SR) approach does not parameters. It calculates the difference between the original signal and a smooth one in the log spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT appro small targets from the reconstruction that is calculated only by the phase spectrum of the in It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computa HFT approach explains the intrinsic theory of saliency detector in the frequency domain and u filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circ e.g., ground-sky background [8], which is common in the helicopter view. In this situation, easily mixed up with the background clutters in size and easily overlapped by vegetation, ro bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by visual system (HVS) has been designed in this paper. HVS is a kind of layered image process consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. this paper is organized as follows. In Section 2, we describe the framework of the proposed for small target detection. In Section 3, we present the experimental results. Section 4 is the of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visu selection mechanism to make it easy to understand and analyze. On the other hand, as a of low-level artificial vision processing, it facilitates subsequent procedures by reducing com cost, which is a key consideration in real-time applications. Based on the above knowledge, a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predete a saliency map(SM) is obtained and the most salient region is picked up to improve detec In detection stage, a support vector machines (SVM) classifier is used to get the target quick Due to the two layers structure, the algorithm computational complexity becomes the prim For the simplicity and high processing speed, saliency detection methods in the frequency ś urf (4) where 015, xx 2 cent years, a series of simple and fast algorithm based on fourier transform was proposed, ral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier T) [7]. With regards to small target detection, frequency domain method is quite different ethods. It transforms the airspace information to the frequency domain, defines significant sts in the frequency domain. While, spectral residual (SR) approach does not rely on the t calculates the difference between the original signal and a smooth one in the log amplitude d then makes up a saliency map by transforming SR to spatial domain. PFT approach detects from the reconstruction that is calculated only by the phase spectrum of the input signal. omputation of SR in the amplitude spectrum, which saves about 1/3 computational cost. h explains the intrinsic theory of saliency detector in the frequency domain and use spectral ess repeated patterns.S numerous methods have been proposed, many of them may fail in certain circumstances, sky background [8], which is common in the helicopter view. In this situation, targets are up with the background clutters in size and easily overlapped by vegetation, roads, rivers, esulting a huge false rate in traditional algorithm. o design an appropriate method, a small target detection method inspired by the human (HVS) has been designed in this paper. HVS is a kind of layered image processing system optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of organized as follows. In Section 2, we describe the framework of the proposed algorithm et detection. In Section 3, we present the experimental results. Section 4 is the conclusion .
Based on Human Visual System rk of the Proposed Two Stage Algorithm-A Brief Description es the scene into small patches and select important information through visual attention chanism to make it easy to understand and analyze. On the other hand, as a component artificial vision processing, it facilitates subsequent procedures by reducing computational s a key consideration in real-time applications. Based on the above knowledge, we propose consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, ap(SM) is obtained and the most salient region is picked up to improve detection speed. tage, a support vector machines (SVM) classifier is used to get the target quickly. two layers structure, the algorithm computational complexity becomes the prime concern. licity and high processing speed, saliency detection methods in the frequency domain are surf is called the "surface-residue dataset" that contains a total of 13,771 surfaces residues, of which 2828 are interfacial residues belonging to the positive subset Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier trans such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hy transform(HFT) [7]. With regards to small target detection, frequency domain meth from other methods. It transforms the airspace information to the frequency domain target and tests in the frequency domain. While, spectral residual (SR) approach d parameters. It calculates the difference between the original signal and a smooth one spectrum, and then makes up a saliency map by transforming SR to spatial domain. P small targets from the reconstruction that is calculated only by the phase spectrum It omits the computation of SR in the amplitude spectrum, which saves about 1/3 HFT approach explains the intrinsic theory of saliency detector in the frequency dom filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in ce e.g., ground-sky background [8], which is common in the helicopter view. In this s easily mixed up with the background clutters in size and easily overlapped by vege bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method ins visual system (HVS) has been designed in this paper. HVS is a kind of layered imag consisting of optical system, retina and visual pathways, which is nonuniform and n this paper is organized as follows. In Section 2, we describe the framework of the for small target detection. In Section 3, we present the experimental results. Section of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information thro selection mechanism to make it easy to understand and analyze. On the other ha of low-level artificial vision processing, it facilitates subsequent procedures by redu cost, which is a key consideration in real-time applications. Based on the above kno a framework consisting of two stages inspired by HVS as follows (See Figure 1). In a saliency map(SM) is obtained and the most salient region is picked up to impro In detection stage, a support vector machines (SVM) classifier is used to get the targ Due to the two layers structure, the algorithm computational complexity become For the simplicity and high processing speed, saliency detection methods in the fre s urf while 10,943 are non-interfacial residues belonging the negative subset Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed. In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Due to the two layers structure, the algorithm computational complexity becomes the prime concern. For the simplicity and high processing speed, saliency detection methods in the frequency domain aré surf , and Ť is the symbol of union in the set theory. If considering all the residues as done in [11], however, the corresponding benchmark dataset can be expressed by: Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed. In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Due to the two layers structure, the algorithm computational complexity becomes the prime concern.
all " Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was propo such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fo transform(HFT) [7]. With regards to small target detection, frequency domain method is quite diffe from other methods. It transforms the airspace information to the frequency domain, defines signifi target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on parameters. It calculates the difference between the original signal and a smooth one in the log ampli spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach de small targets from the reconstruction that is calculated only by the phase spectrum of the input sig It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spe filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstan e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, riv bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the hu visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing sy consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The re this paper is organized as follows. In Section 2, we describe the framework of the proposed algor for small target detection. In Section 3, we present the experimental results. Section 4 is the conclu of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual atten selection mechanism to make it easy to understand and analyze. On the other hand, as a compo of low-level artificial vision processing, it facilitates subsequent procedures by reducing computati cost, which is a key consideration in real-time applications. Based on the above knowledge, we pro a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection st a saliency map(SM) is obtained and the most salient region is picked up to improve detection sp In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Due to the two layers structure, the algorithm computational complexity becomes the prime conc à ll ď Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomp transform(HFT) [7]. With regards to small target detection, frequency domain method is qui from other methods. It transforms the airspace information to the frequency domain, defines target and tests in the frequency domain. While, spectral residual (SR) approach does not parameters. It calculates the difference between the original signal and a smooth one in the log spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT appro small targets from the reconstruction that is calculated only by the phase spectrum of the in It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computa HFT approach explains the intrinsic theory of saliency detector in the frequency domain and u filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circ e.g., ground-sky background [8], which is common in the helicopter view. In this situation, easily mixed up with the background clutters in size and easily overlapped by vegetation, ro bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by visual system (HVS) has been designed in this paper. HVS is a kind of layered image process consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. this paper is organized as follows. In Section 2, we describe the framework of the proposed for small target detection. In Section 3, we present the experimental results. Section 4 is the of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visu selection mechanism to make it easy to understand and analyze. On the other hand, as a of low-level artificial vision processing, it facilitates subsequent procedures by reducing com cost, which is a key consideration in real-time applications. Based on the above knowledge, w a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetec a saliency map(SM) is obtained and the most salient region is picked up to improve detec In detection stage, a support vector machines (SVM) classifier is used to get the target quickl Due to the two layers structure, the algorithm computational complexity becomes the prim á ll (5) where 015, xx 2 cent years, a series of simple and fast algorithm based on fourier transform was proposed, ral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier T) [7]. With regards to small target detection, frequency domain method is quite different ethods. It transforms the airspace information to the frequency domain, defines significant ts in the frequency domain. While, spectral residual (SR) approach does not rely on the t calculates the difference between the original signal and a smooth one in the log amplitude then makes up a saliency map by transforming SR to spatial domain. PFT approach detects from the reconstruction that is calculated only by the phase spectrum of the input signal. omputation of SR in the amplitude spectrum, which saves about 1/3 computational cost. h explains the intrinsic theory of saliency detector in the frequency domain and use spectral ess repeated patterns.S numerous methods have been proposed, many of them may fail in certain circumstances, sky background [8], which is common in the helicopter view. In this situation, targets are up with the background clutters in size and easily overlapped by vegetation, roads, rivers, esulting a huge false rate in traditional algorithm. o design an appropriate method, a small target detection method inspired by the human (HVS) has been designed in this paper. HVS is a kind of layered image processing system optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of organized as follows. In Section 2, we describe the framework of the proposed algorithm et detection. In Section 3, we present the experimental results. Section 4 is the conclusion .
Based on Human Visual System rk of the Proposed Two Stage Algorithm-A Brief Description es the scene into small patches and select important information through visual attention chanism to make it easy to understand and analyze. On the other hand, as a component artificial vision processing, it facilitates subsequent procedures by reducing computational s a key consideration in real-time applications. Based on the above knowledge, we propose consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, ap(SM) is obtained and the most salient region is picked up to improve detection speed. tage, a support vector machines (SVM) classifier is used to get the target quickly. all is called the "all-residue dataset" that contains a total of 27,442 residues, of which 2828 are interfacial residues belonging to the positive subset Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercompl transform(HFT) [7]. With regards to small target detection, frequency domain method is quit from other methods. It transforms the airspace information to the frequency domain, defines target and tests in the frequency domain. While, spectral residual (SR) approach does not r parameters. It calculates the difference between the original signal and a smooth one in the log spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approa small targets from the reconstruction that is calculated only by the phase spectrum of the inp It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computat HFT approach explains the intrinsic theory of saliency detector in the frequency domain and u filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circu e.g., ground-sky background [8], which is common in the helicopter view. In this situation, easily mixed up with the background clutters in size and easily overlapped by vegetation, roa bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by t visual system (HVS) has been designed in this paper. HVS is a kind of layered image processi consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. T this paper is organized as follows. In Section 2, we describe the framework of the proposed for small target detection. In Section 3, we present the experimental results. Section 4 is the c of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visua selection mechanism to make it easy to understand and analyze. On the other hand, as a c of low-level artificial vision processing, it facilitates subsequent procedures by reducing com cost, which is a key consideration in real-time applications. Based on the above knowledge, w a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetec a saliency map(SM) is obtained and the most salient region is picked up to improve detect In detection stage, a support vector machines (SVM) classifier is used to get the target quickly  clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost.
HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed. In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. For readers' convenience, given in S1 Dataset (List of the 99 proteins and their residues' attributions associated with the protein-protein binding sites is in Supplementary Materials) is a combination of the two benchmark datasets, where those labeled in column 3 are all the residues determined by experiments, those in column 4 are of surface and non-surface residues, and those in column 5 are of interface and non-interface residues.
As pointed out in a comprehensive review [27] there is no need to separate a benchmark dataset into a training dataset and a testing dataset for examining the quality of a prediction method if it is tested by the jackknife test or subsampling (K-fold) cross-validation test because the outcome obtained via this kind of approach is actually from a combination of many different independent dataset tests.

Flexible Sliding Window Approach
Given a protein chain as expressed in Equation (1), the sliding window approach [28] and flexible sliding window approach [29] are often used to investigate its various posttranslational modification (PTM) sites [16,[30][31][32][33][34] and HIV (human immunodeficiency virus) protease cleavage sites [35]. Here, we also use it to study protein-protein binding sites. In the sliding window approach, a scaled window is denoted by r-ξ,`ξs [28], and its width is 2ξ`1, where ξ is an integer. When sliding it along a protein chain P, one can see through the window a series of consecutive peptide segments as formulated by: clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Algorithm Based on Human Visual System
Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercompl transform(HFT) [7]. With regards to small target detection, frequency domain method is quit from other methods. It transforms the airspace information to the frequency domain, defines target and tests in the frequency domain. While, spectral residual (SR) approach does not r parameters. It calculates the difference between the original signal and a smooth one in the log spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approa small targets from the reconstruction that is calculated only by the phase spectrum of the inp It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computat HFT approach explains the intrinsic theory of saliency detector in the frequency domain and u filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circu e.g., ground-sky background [8], which is common in the helicopter view. In this situation, easily mixed up with the background clutters in size and easily overlapped by vegetation, roa bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by t visual system (HVS) has been designed in this paper. HVS is a kind of layered image processi consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. T this paper is organized as follows. In Section 2, we describe the framework of the proposed for small target detection. In Section 3, we present the experimental results. Section 4 is the c of this article.

Algorithm Based on Human Visual System
where R´ξ represents the ξ-th upstream amino acid residue from the center, R`ξ the ξ-th downstream amino acid residue, and so forth. The amino acid residue Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hyperc transform(HFT) [7]. With regards to small target detection, frequency domain method i from other methods. It transforms the airspace information to the frequency domain, de target and tests in the frequency domain. While, spectral residual (SR) approach does parameters. It calculates the difference between the original signal and a smooth one in th spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT a small targets from the reconstruction that is calculated only by the phase spectrum of t It omits the computation of SR in the amplitude spectrum, which saves about 1/3 com HFT approach explains the intrinsic theory of saliency detector in the frequency domain filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain e.g., ground-sky background [8], which is common in the helicopter view. In this situa easily mixed up with the background clutters in size and easily overlapped by vegetatio bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired visual system (HVS) has been designed in this paper. HVS is a kind of layered image pro consisting of optical system, retina and visual pathways, which is nonuniform and nonlin this paper is organized as follows. In Section 2, we describe the framework of the prop for small target detection. In Section 3, we present the experimental results. Section 4 is of this article. 0 at the center is the targeted residue. When its sequence position in P (cf. Equation (1) is less than ξ or greater L´ξ, the corresponding P ξ` 5, xx 2 nt years, a series of simple and fast algorithm based on fourier transform was proposed, l residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier ) [7]. With regards to small target detection, frequency domain method is quite different hods. It transforms the airspace information to the frequency domain, defines significant in the frequency domain. While, spectral residual (SR) approach does not rely on the alculates the difference between the original signal and a smooth one in the log amplitude hen makes up a saliency map by transforming SR to spatial domain. PFT approach detects om the reconstruction that is calculated only by the phase spectrum of the input signal. putation of SR in the amplitude spectrum, which saves about 1/3 computational cost. explains the intrinsic theory of saliency detector in the frequency domain and use spectral s repeated patterns.R merous methods have been proposed, many of them may fail in certain circumstances, y background [8], which is common in the helicopter view. In this situation, targets are with the background clutters in size and easily overlapped by vegetation, roads, rivers, ulting a huge false rate in traditional algorithm. design an appropriate method, a small target detection method inspired by the human VS) has been designed in this paper. HVS is a kind of layered image processing system tical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of ganized as follows. In Section 2, we describe the framework of the proposed algorithm 0˘i s defined, rather than by P of Equation (1), but by the following dummy protein chain: Ppdummyq " Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of ξ¨¨Ä lgorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of 2 Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost.
HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system L Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourie transform(HFT) [7]. With regards to small target detection, frequency domain method is quite differen from other methods. It transforms the airspace information to the frequency domain, defines significan target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on th parameters. It calculates the difference between the original signal and a smooth one in the log amplitud spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detect small targets from the reconstruction that is calculated only by the phase spectrum of the input signal It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectra filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets ar easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system L´1¨¨Ä lgorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was p such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomple transform(HFT) [7]. With regards to small target detection, frequency domain method is quite from other methods. It transforms the airspace information to the frequency domain, defines si target and tests in the frequency domain. While, spectral residual (SR) approach does not re parameters. It calculates the difference between the original signal and a smooth one in the log a spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approac small targets from the reconstruction that is calculated only by the phase spectrum of the inp It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computatio HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circum e.g., ground-sky background [8], which is common in the helicopter view. In this situation, ta easily mixed up with the background clutters in size and easily overlapped by vegetation, road bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by th visual system (HVS) has been designed in this paper. HVS is a kind of layered image processin where the symbol õ stands for a mirror, the dummy segment Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier trans such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], h transform(HFT) [7]. With regards to small target detection, frequency domain meth from other methods. It transforms the airspace information to the frequency domai target and tests in the frequency domain. While, spectral residual (SR) approach parameters. It calculates the difference between the original signal and a smooth one spectrum, and then makes up a saliency map by transforming SR to spatial domain. P small targets from the reconstruction that is calculated only by the phase spectrum It omits the computation of SR in the amplitude spectrum, which saves about 1/3 HFT approach explains the intrinsic theory of saliency detector in the frequency dom filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in ce e.g., ground-sky background [8], which is common in the helicopter view. In this easily mixed up with the background clutters in size and easily overlapped by vege bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method ins ξ¨¨Ä lgorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourie such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) transform(HFT) [7]. With regards to small target detection, frequency domai from other methods. It transforms the airspace information to the frequency d target and tests in the frequency domain. While, spectral residual (SR) app parameters. It calculates the difference between the original signal and a smoo spectrum, and then makes up a saliency map by transforming SR to spatial dom small targets from the reconstruction that is calculated only by the phase sp It omits the computation of SR in the amplitude spectrum, which saves abo HFT approach explains the intrinsic theory of saliency detector in the frequen filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fai e.g., ground-sky background [8], which is common in the helicopter view. I easily mixed up with the background clutters in size and easily overlapped b bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection meth 2 Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fou such as spectral residual(SR) [5], phase spectrum of fourier transform(PF transform(HFT) [7]. With regards to small target detection, frequency dom from other methods. It transforms the airspace information to the frequenc target and tests in the frequency domain. While, spectral residual (SR) a parameters. It calculates the difference between the original signal and a sm spectrum, and then makes up a saliency map by transforming SR to spatial d small targets from the reconstruction that is calculated only by the phase It omits the computation of SR in the amplitude spectrum, which saves a HFT approach explains the intrinsic theory of saliency detector in the frequ filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may e.g., ground-sky background [8], which is common in the helicopter view easily mixed up with the background clutters in size and easily overlapped bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection me 1 stands for the image of R 1 R 2¨¨¨Rξ reflected by the mirror, and the dummy segment Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier trans such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], h transform(HFT) [7]. With regards to small target detection, frequency domain meth from other methods. It transforms the airspace information to the frequency domai target and tests in the frequency domain. While, spectral residual (SR) approach parameters. It calculates the difference between the original signal and a smooth one spectrum, and then makes up a saliency map by transforming SR to spatial domain. P small targets from the reconstruction that is calculated only by the phase spectrum It omits the computation of SR in the amplitude spectrum, which saves about 1/3 HFT approach explains the intrinsic theory of saliency detector in the frequency dom filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in ce e.g., ground-sky background [8], which is common in the helicopter view. In this easily mixed up with the background clutters in size and easily overlapped by vege bridges [9], resulting a huge false rate in traditional algorithm. L Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier tr such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6] transform(HFT) [7]. With regards to small target detection, frequency domain m from other methods. It transforms the airspace information to the frequency dom target and tests in the frequency domain. While, spectral residual (SR) approa parameters. It calculates the difference between the original signal and a smooth o spectrum, and then makes up a saliency map by transforming SR to spatial domain small targets from the reconstruction that is calculated only by the phase spectr It omits the computation of SR in the amplitude spectrum, which saves about 1 HFT approach explains the intrinsic theory of saliency detector in the frequency d filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in e.g., ground-sky background [8], which is common in the helicopter view. In th easily mixed up with the background clutters in size and easily overlapped by v bridges [9], resulting a huge false rate in traditional algorithm.

L´1¨¨Ä
lgorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on f such as spectral residual(SR) [5], phase spectrum of fourier transform(P transform(HFT) [7]. With regards to small target detection, frequency d from other methods. It transforms the airspace information to the freque target and tests in the frequency domain. While, spectral residual (SR) parameters. It calculates the difference between the original signal and a spectrum, and then makes up a saliency map by transforming SR to spatia small targets from the reconstruction that is calculated only by the phas It omits the computation of SR in the amplitude spectrum, which saves HFT approach explains the intrinsic theory of saliency detector in the fre filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them ma e.g., ground-sky background [8], which is common in the helicopter vie easily mixed up with the background clutters in size and easily overlapp bridges [9], resulting a huge false rate in traditional algorithm. image of R L´ξ`1¨¨¨RL´1 R L ( Figure 1). Accordingly, P(dummy) of Equation (7) is also called the mirror-extended chain of protein P.
Thus, for each of the L amino acid residues in protein P, we have a working protein segment as defined by Equation (6). In the current study, the p2ξ`1q-tuple peptide P ξ` transform(HFT) [7]. With regards to small target detection, frequenc from other methods. It transforms the airspace information to the fre target and tests in the frequency domain. While, spectral residual ( parameters. It calculates the difference between the original signal an spectrum, and then makes up a saliency map by transforming SR to sp small targets from the reconstruction that is calculated only by the p It omits the computation of SR in the amplitude spectrum, which s HFT approach explains the intrinsic theory of saliency detector in the filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them e.g., ground-sky background [8], which is common in the helicopter easily mixed up with the background clutters in size and easily over bridges [9], resulting a huge false rate in traditional algorithm. In order to design an appropriate method, a small target detecti visual system (HVS) has been designed in this paper. HVS is a kind o consisting of optical system, retina and visual pathways, which is non this paper is organized as follows. In Section 2, we describe the fra for small target detection. In Section 3, we present the experimental of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Desc
HVS divides the scene into small patches and select important in selection mechanism to make it easy to understand and analyze. O of low-level artificial vision processing, it facilitates subsequent pro cost, which is a key consideration in real-time applications. Based on a framework consisting of two stages inspired by HVS as follows (S a saliency map(SM) is obtained and the most salient region is pick In detection stage, a support vector machines (SVM) classifier is use Due to the two layers structure, the algorithm computational comp For the simplicity and high processing speed, saliency detection me 0˘c an be further classified into the following categories: clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed. In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Due to the two layers structure, the algorithm computational complexity becomes the prime concern. For the simplicity and high processing speed, saliency detection methods in the frequency domain are clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourie transform(HFT) [7]. With regards to small target detection, frequency domain method is quite differen from other methods. It transforms the airspace information to the frequency domain, defines significan target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detect small targets from the reconstruction that is calculated only by the phase spectrum of the input signal It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectra filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest o this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a componen of low-level artificial vision processing, it facilitates subsequent procedures by reducing computationa cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Due to the two layers structure, the algorithm computational complexity becomes the prime concern For the simplicity and high processing speed, saliency detection methods in the frequency domain are 0˘, if its center is a PPBS Pξ` clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourie transform(HFT) [7]. With regards to small target detection, frequency domain method is quite differen from other methods. It transforms the airspace information to the frequency domain, defines significan target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectra filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest o this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a componen of low-level artificial vision processing, it facilitates subsequent procedures by reducing computationa cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Due to the two layers structure, the algorithm computational complexity becomes the prime concern For the simplicity and high processing speed, saliency detection methods in the frequency domain are 0˘, otherwise (8) where P represents "a member of" in the set theory. Figure 1. A schematic drawing to show how to use the extended chain of Equation (7) to define the working segments of Equation (6) for those sites when their sequence positions in the protein are less than ξ or greater L´ξ , where the left dummy segment stands for the mirror image of R 1 R 2¨¨¨Rξ at N-terminus and the right dummy segment for that of R L´ξ`1¨¨¨RL´1 R L at the C-terminus.

Using Pseudo Amino Acid Composition to Represent Peptide Chains
One of the most challenging problems in computational biology today is how to effectively formulate the sequence of a biological sample (such as protein/peptide, and DNA/RNA) with a discrete model or a vector that can considerably keep its sequence order information or capture its key features. The reasons are as follows: (1) If using the sequential model, i.e., the model in which all the samples are represented by their original sequences, it is hardly able to train a machine that can cover all the possible cases concerned, as elaborated in [36]; (2) All the existing computational algorithms, such as optimization approach [37], correlation-angle approach [38], covariance discriminant (CD) [39], neural network [40], K-nearest neighbor (KNN) [41], OET-KNN [42], SLLE algorithm [43], random forest [44], Fuzzy K-nearest neighbor [45], and ML-KNN algorithm [46], can only handle vector but not sequence samples.
However, a vector defined in a discrete model may completely lose the sequence-order information as elaborated in [47,48]. To cope with such a dilemma, the approach of pseudo amino acid composition [36,49] or Chou's PseAAC [50,51] was proposed. Ever since it was introduced in 2001 [36], the concept of PseAAC has been penetrating into nearly all the areas of computational biology (see, e.g., [52][53][54][55][56] as well as a long list of references cited in [48,57] and a recent review [58]). It has also been selected as a special topic for a special issue on "drug development and biomedicine" [59]. Recently, the concept of PseAAC was further extended to represent the feature vectors of DNA and nucleotides [60][61][62][63][64]. Because of its being widely and increasingly used, three types of open access soft-ware, called "PseAAC-Builder" [65], "propy" [50], and "PseAAC-General" [57], were established: the former two are for generating various modes of Chou's special PseAAC; while the 3rd one for those of Chou's general PseAAC.
According to [12], PseAAC can be generally formulated as: where T is the transpose operator, while Ω an integer to reflect the vector's dimension. The value of Ω as well as the components Ψ u " pu " 1, 2,¨¨¨, Ωq in Equation (9) will depend on how to extract the desired information from a peptide sequence. Below, we are to describe how to extract the useful information from the aforementioned benchmark datasets (cf. Equations (4) and (5)) to define the working protein segments via Equation (9). For the convenience of formulation below, we convert the p2ξ`1q-tuple peptide in Equation (6) to:

Amino Acid Code
Physicochemical Property (cf. Equation (11) Thus, the peptide segment P ξ of Equation (10) can be encoded into seven different numerical series, as formulated by: where Φ p1q 1 is the hydrophobicity value of R 1 in Equation (10), Φ p2q 2 the hydrophilicity value of R 2 , and so forth. Note that before substituting the physicochemical values of Table 2 into Equation (10), they all are subjected to the following standard conversion: where the symbol x y means taking the average for the quantity therein over the 20 amino acid types, and SD means the corresponding standard deviation. The converted values via Equation (12) will have zero mean value over the 20 amino acid types, and will remain unchanged if they go through the same standard conversion procedure again.

Stationary Wavelet Transform Approach
The low-frequency internal motion is a very important feature of biomacromolecules (see, e.g., [73][74][75]. Many marvelous biological functions in proteins and DNA and their profound dynamic mechanisms, such as switch between active and inactive states [76,77], cooperative effects [78], allosteric transition [79][80][81], intercalation of drugs into DNA [82], and assembly of microtubules [83], can be revealed by studying their low-frequency internal motions as summarized in a comprehensive review [84]. Low frequency Fourier spectrum was also used by Liu et al. [85] to develop a sequence-based method for predicting membrane protein types. In view of this, it would be intriguing to introduce the stationary wavelet transform into the current study. The stationary wavelet transform (SWT) [86] is a wavelet transform algorithm designed to overcome the lack of shift-invariance of the discrete wavelet transform (DWT) [87]. Shift-invariance is achieved by removing the downsamplers and upsamplers in the DWT and upsampling (insert zero) the filter coefficients by a factor of 2 j´1 in the j-th level of the algorithm. The SWT is an inherently redundant scheme as the output of each level of SWT contains the same number of samples as the input-so for a decomposition of N levels there is a redundancy of N in the wavelet coefficients. Shown in Figure 2 is the block diagram depicting the digital implementation of SWT. As we can see from the figure, the input peptide segment is decomposed recursively in the low-frequency part.
The concrete procedure of using the SWT to denote the p2ξ`1q-tuple peptides is as follows. For each of the p2ξ`1q-tuple peptides generated by sliding the scaled windowr´ξ,`ξs along the protein chain concerned, the SWT was used to decompose it based on the amino acid values encoded by the seven physicochemical properties as given in Equation (11). Daubechies of number 1 (Db1) wavelet was selected because its wavelet possesses a lower vanish moment and easily generates non-zero coefficients for the ensemble learning framework that will be introduced later.
Preliminary tests indicated that, when ξ " 7, i.e., the working segments are 15-tuple peptides, the outcomes thus obtained were most promising. Accordingly, we only consider the case of ξ " 7 hereafter. Using the SWT approach, we have generated five sub-bands (Figure 2), each of which has four coefficients: (1) α i , the maximum of the wavelet coefficients in the sub-band i p1, 2,¨¨¨5q; (2) β i , the corresponding mean of the wavelet coefficients; (3) γ i , the corresponding minimum of the wavelet coefficients; (4) δ i , the corresponding standard deviation of the wavelet coefficients. Therefore, for each working segment, we can get a feature vector that contains Ω " 5ˆ4 " 20 components by using each of the seven physicochemical properties of Equation (11). In other words, we have seven different modes of PseAAC as given below: where:

Optimizing Imbalanced Training Datasets
In the current benchmark dataset Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost.
HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was propose such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fouri transform(HFT) [7]. With regards to small target detection, frequency domain method is quite differe from other methods. It transforms the airspace information to the frequency domain, defines significa target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on th parameters. It calculates the difference between the original signal and a smooth one in the log amplitud spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detec small targets from the reconstruction that is calculated only by the phase spectrum of the input signa It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational co HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectr filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstance e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets a easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, river bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the huma visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing syste consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest this paper is organized as follows. In Section 2, we describe the framework of the proposed algorith for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusio of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attentio selection mechanism to make it easy to understand and analyze. On the other hand, as a compone of low-level artificial vision processing, it facilitates subsequent procedures by reducing computation cost, which is a key consideration in real-time applications. Based on the above knowledge, we propo a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stag a saliency map(SM) is obtained and the most salient region is picked up to improve detection spee In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. clutters. In recent years, a series of simple and fast algorithm based on fourier t such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6 transform(HFT) [7]. With regards to small target detection, frequency domain m from other methods. It transforms the airspace information to the frequency dom target and tests in the frequency domain. While, spectral residual (SR) approa parameters. It calculates the difference between the original signal and a smooth spectrum, and then makes up a saliency map by transforming SR to spatial domai small targets from the reconstruction that is calculated only by the phase spect It omits the computation of SR in the amplitude spectrum, which saves about HFT approach explains the intrinsic theory of saliency detector in the frequency filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in e.g., ground-sky background [8], which is common in the helicopter view. In t easily mixed up with the background clutters in size and easily overlapped by v bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method visual system (HVS) has been designed in this paper. HVS is a kind of layered im consisting of optical system, retina and visual pathways, which is nonuniform an this paper is organized as follows. In Section 2, we describe the framework of for small target detection. In Section 3, we present the experimental results. Sec of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information selection mechanism to make it easy to understand and analyze. On the othe of low-level artificial vision processing, it facilitates subsequent procedures by cost, which is a key consideration in real-time applications. Based on the above a framework consisting of two stages inspired by HVS as follows (See Figure 1 a saliency map(SM) is obtained and the most salient region is picked up to im In detection stage, a support vector machines (SVM) classifier is used to get the  Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fo such as spectral residual(SR) [5], phase spectrum of fourier transform(PF transform(HFT) [7]. With regards to small target detection, frequency dom from other methods. It transforms the airspace information to the frequen target and tests in the frequency domain. While, spectral residual (SR) a parameters. It calculates the difference between the original signal and a sm spectrum, and then makes up a saliency map by transforming SR to spatial small targets from the reconstruction that is calculated only by the phase It omits the computation of SR in the amplitude spectrum, which saves HFT approach explains the intrinsic theory of saliency detector in the frequ filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may e.g., ground-sky background [8], which is common in the helicopter view easily mixed up with the background clutters in size and easily overlappe bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection m visual system (HVS) has been designed in this paper. HVS is a kind of lay consisting of optical system, retina and visual pathways, which is nonunifo this paper is organized as follows. In Section 2, we describe the framewo for small target detection. In Section 3, we present the experimental resul of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Descriptio
HVS divides the scene into small patches and select important inform selection mechanism to make it easy to understand and analyze. On the of low-level artificial vision processing, it facilitates subsequent procedur cost, which is a key consideration in real-time applications. Based on the a framework consisting of two stages inspired by HVS as follows (See Fi a saliency map(SM) is obtained and the most salient region is picked up In detection stage, a support vector machines (SVM) classifier is used to g clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost.
HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly.  clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost.
HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly.  utters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, ch as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier nsform(HFT) [7]. With regards to small target detection, frequency domain method is quite different m other methods. It transforms the airspace information to the frequency domain, defines significant rget and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the rameters. It calculates the difference between the original signal and a smooth one in the log amplitude ectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects all targets from the reconstruction that is calculated only by the phase spectrum of the input signal. omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. FT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral ter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are sily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, idges [9], resulting a huge false rate in traditional algorithm. In order to design an appropriate method, a small target detection method inspired by the human sual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system nsisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of is paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm r small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention lection mechanism to make it easy to understand and analyze. On the other hand, as a component low-level artificial vision processing, it facilitates subsequent procedures by reducing computational st, which is a key consideration in real-time applications. Based on the above knowledge, we propose framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, saliency map(SM) is obtained and the most salient region is picked up to improve detection speed. detection stage, a support vector machines (SVM) classifier is used to get the target quickly. clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost.
HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. clutters. In recent years, a series of simple and fast algorithm based on fourier transform was propo such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fou transform(HFT) [7]. With regards to small target detection, frequency domain method is quite diffe from other methods. It transforms the airspace information to the frequency domain, defines signifi target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on parameters. It calculates the difference between the original signal and a smooth one in the log ampli spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach de small targets from the reconstruction that is calculated only by the phase spectrum of the input sig It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational c HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spe filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstan e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, riv bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the hu visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing sys consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The re this paper is organized as follows. In Section 2, we describe the framework of the proposed algor for small target detection. In Section 3, we present the experimental results. Section 4 is the conclu of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual atten selection mechanism to make it easy to understand and analyze. On the other hand, as a compo of low-level artificial vision processing, it facilitates subsequent procedures by reducing computati cost, which is a key consideration in real-time applications. Based on the above knowledge, we pro a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection st a saliency map(SM) is obtained and the most salient region is picked up to improve detection sp In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. utters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, ch as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier nsform(HFT) [7]. With regards to small target detection, frequency domain method is quite different m other methods. It transforms the airspace information to the frequency domain, defines significant rget and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the rameters. It calculates the difference between the original signal and a smooth one in the log amplitude ectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects all targets from the reconstruction that is calculated only by the phase spectrum of the input signal. omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. FT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral ter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are sily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, idges [9], resulting a huge false rate in traditional algorithm. In order to design an appropriate method, a small target detection method inspired by the human sual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system nsisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of is paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm r small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention lection mechanism to make it easy to understand and analyze. On the other hand, as a component low-level artificial vision processing, it facilitates subsequent procedures by reducing computational st, which is a key consideration in real-time applications. Based on the above knowledge, we propose framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, saliency map(SM) is obtained and the most salient region is picked up to improve detection speed. detection stage, a support vector machines (SVM) classifier is used to get the target quickly. clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed. In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. clutters. In recent years, a series of simple and fast algorithm based on fourier transform was propose such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fouri transform(HFT) [7]. With regards to small target detection, frequency domain method is quite differe from other methods. It transforms the airspace information to the frequency domain, defines significa target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on t parameters. It calculates the difference between the original signal and a smooth one in the log amplitu spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detec small targets from the reconstruction that is calculated only by the phase spectrum of the input sign It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational co HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectr filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstance e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets a easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rive bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the hum visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing syste consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest this paper is organized as follows. In Section 2, we describe the framework of the proposed algorith for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusi of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attenti selection mechanism to make it easy to understand and analyze. On the other hand, as a compone of low-level artificial vision processing, it facilitates subsequent procedures by reducing computation cost, which is a key consideration in real-time applications. Based on the above knowledge, we propo a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stag a saliency map(SM) is obtained and the most salient region is picked up to improve detection spee In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. for all residues (15) where the figures in the parentheses denote the sample numbers taken from Section 2.1. As we can see from Equation (15), the numbers of the negative samples are nearly nine and four times the sizes of the corresponding positive samples for the all-residue and surface-residue benchmark datasets, respectively. Although this might reflect the real world in which the non-binding sites are always the majority compared with the binding ones, a predictor trained by such a highly skewed benchmark dataset would inevitably have the bias consequence that many binding sites might be mispredicted as non-binding ones [88]. Actually, what is really the most intriguing information for us is the information about the binding sites. Therefore, it is important to find an effective approach to optimize the unbalanced training dataset and minimize this kind of bias consequence. To realize this, we took the following procedures.
First, we used the K-Nearest Neighbors Cleaning (KNNC) treatment to remove some redundant negative samples from the negative subset so as to reduce its statistical noise. The detailed process can be described below: (i) for each of the samples in the negative subset Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier t such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6 transform(HFT) [7]. With regards to small target detection, frequency domain m from other methods. It transforms the airspace information to the frequency dom target and tests in the frequency domain. While, spectral residual (SR) approa parameters. It calculates the difference between the original signal and a smooth spectrum, and then makes up a saliency map by transforming SR to spatial domai small targets from the reconstruction that is calculated only by the phase spect It omits the computation of SR in the amplitude spectrum, which saves about HFT approach explains the intrinsic theory of saliency detector in the frequency filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in e.g., ground-sky background [8], which is common in the helicopter view. In t easily mixed up with the background clutters in size and easily overlapped by v bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method visual system (HVS) has been designed in this paper. HVS is a kind of layered im consisting of optical system, retina and visual pathways, which is nonuniform an this paper is organized as follows. In Section 2, we describe the framework of for small target detection. In Section 3, we present the experimental results. Sec of this article. where K may be any integer (such as 3 or 8), and its final value will be discussed later; (ii) if one of its K nearest neighbors belongs to the positive subset Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was prop such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fo transform(HFT) [7]. With regards to small target detection, frequency domain method is quite dif from other methods. It transforms the airspace information to the frequency domain, defines signi target and tests in the frequency domain. While, spectral residual (SR) approach does not rely o parameters. It calculates the difference between the original signal and a smooth one in the log amp spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach d small targets from the reconstruction that is calculated only by the phase spectrum of the input s It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use sp filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumsta e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targe easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, r bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the h visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing sy consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The r this paper is organized as follows. In Section 2, we describe the framework of the proposed algo for small target detection. In Section 3, we present the experimental results. Section 4 is the concl of this article.

Algorithm Based on Human Visual System
`, remove the negative sample from Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm b such as spectral residual(SR) [5], phase spectrum of fourier tra transform(HFT) [7]. With regards to small target detection, freq from other methods. It transforms the airspace information to t target and tests in the frequency domain. While, spectral resid parameters. It calculates the difference between the original sign spectrum, and then makes up a saliency map by transforming SR small targets from the reconstruction that is calculated only by It omits the computation of SR in the amplitude spectrum, wh HFT approach explains the intrinsic theory of saliency detector filter to suppress repeated patterns.S Although numerous methods have been proposed, many of e.g., ground-sky background [8], which is common in the helic easily mixed up with the background clutters in size and easily bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target d visual system (HVS) has been designed in this paper. HVS is a consisting of optical system, retina and visual pathways, which this paper is organized as follows. In Section 2, we describe th for small target detection. In Section 3, we present the experim of this article.  [90], and Liu et al. [91] although their details are different with the current practice. Also, the current KNNC approach is more flexible because it contains a variable K and hence can be used to deal with various different training datasets.

Algorithm Based on Human Visual System
Second, we used the Inserting Hypothetical Training Samples (IHTS) treatment to add some hypothetical positive samples into the positive subset so as to enhance the ability in identifying the interactive pairs. For the details of how to generate the hypothetical training samples, see the Monte Calo samples expanding approach in [92,93], or seed-propagation approach in [94], or the SMOTE (synthetic minority over-sampling technique) approach in [95].
After the above two treatments, we can change an original highly skewed training dataset to a balanced training dataset with its positive subset and negative subset having exactly the same size.
It is instructive to point out that the hypothetical samples generated via the IHTS treatment can only be expressed by their feature vectors as defined in Equation (13), but not the real peptide segment samples as given by Equations (6) or (10). Nevertheless, it would be perfectly reasonable to do so because the data directly used to train a predictor were actually the samples' feature vectors but not their sequence codes. This is the key to optimize an imbalanced benchmark dataset in the current study, and the rationale of such an interesting approach will be further elucidated later.

Fusing Multiple Physicochemical Properties
The random forest (RF) algorithm is a powerful algorithm, which has been used in many areas of computational biology (see, e.g., [44,96,97]). The detailed procedures and formulation of RF have been very clearly described in [98], and hence there is no need to repeat here.
As shown in Equations (11)-(13), a peptide segment concerned in the current study can be formulated with seven different PseAAC modes, each of which can be used to train the random forest predictor after the KNNC and IHTS procedures. Accordingly, we have a total of seven individual predictors for identifying PPBS, as formulated by:

PPBS individual predictor "
Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was p such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex transform(HFT) [7]. With regards to small target detection, frequency domain method is quite from other methods. It transforms the airspace information to the frequency domain, defines sig target and tests in the frequency domain. While, spectral residual (SR) approach does not rel parameters. It calculates the difference between the original signal and a smooth one in the log am spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approac small targets from the reconstruction that is calculated only by the phase spectrum of the inpu It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computatio HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circum e.g., ground-sky background [8], which is common in the helicopter view. In this situation, ta easily mixed up with the background clutters in size and easily overlapped by vegetation, road bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. Th this paper is organized as follows. In Section 2, we describe the framework of the proposed a for small target detection. In Section 3, we present the experimental results. Section 4 is the co of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual a selection mechanism to make it easy to understand and analyze. On the other hand, as a com of low-level artificial vision processing, it facilitates subsequent procedures by reducing compu cost, which is a key consideration in real-time applications. Based on the above knowledge, we a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetectio a saliency map(SM) is obtained and the most salient region is picked up to improve detectio In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Due to the two layers structure, the algorithm computational complexity becomes the prime For the simplicity and high processing speed, saliency detection methods in the frequency dom Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercompl transform(HFT) [7]. With regards to small target detection, frequency domain method is quit from other methods. It transforms the airspace information to the frequency domain, defines target and tests in the frequency domain. While, spectral residual (SR) approach does not r parameters. It calculates the difference between the original signal and a smooth one in the log spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approa small targets from the reconstruction that is calculated only by the phase spectrum of the inp It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computat HFT approach explains the intrinsic theory of saliency detector in the frequency domain and u filter to suppress repeated patterns.F Although numerous methods have been proposed, many of them may fail in certain circu e.g., ground-sky background [8], which is common in the helicopter view. In this situation, easily mixed up with the background clutters in size and easily overlapped by vegetation, roa bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by t visual system (HVS) has been designed in this paper. HVS is a kind of layered image processi consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. T this paper is organized as follows. In Section 2, we describe the framework of the proposed for small target detection. In Section 3, we present the experimental results. Section 4 is the c of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visua selection mechanism to make it easy to understand and analyze. On the other hand, as a c of low-level artificial vision processing, it facilitates subsequent procedures by reducing com cost, which is a key consideration in real-time applications. Based on the above knowledge, w a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetec a saliency map(SM) is obtained and the most salient region is picked up to improve detect In detection stage, a support vector machines (SVM) classifier is used to get the target quickly Due to the two layers structure, the algorithm computational complexity becomes the prim For the simplicity and high processing speed, saliency detection methods in the frequency d pkq pk " 1, 2,¨¨¨, 7q (16) where 015, xx 2 ecent years, a series of simple and fast algorithm based on fourier transform was proposed, tral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier FT) [7]. With regards to small target detection, frequency domain method is quite different ethods. It transforms the airspace information to the frequency domain, defines significant sts in the frequency domain. While, spectral residual (SR) approach does not rely on the It calculates the difference between the original signal and a smooth one in the log amplitude d then makes up a saliency map by transforming SR to spatial domain. PFT approach detects from the reconstruction that is calculated only by the phase spectrum of the input signal. computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. h explains the intrinsic theory of saliency detector in the frequency domain and use spectral ress repeated patterns.R numerous methods have been proposed, many of them may fail in certain circumstances, sky background [8], which is common in the helicopter view. In this situation, targets are up with the background clutters in size and easily overlapped by vegetation, roads, rivers, esulting a huge false rate in traditional algorithm. o design an appropriate method, a small target detection method inspired by the human (HVS) has been designed in this paper. HVS is a kind of layered image processing system optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of organized as follows. In Section 2, we describe the framework of the proposed algorithm et detection. In Section 3, we present the experimental results. Section 4 is the conclusion .

ork of the Proposed Two Stage Algorithm-A Brief Description
des the scene into small patches and select important information through visual attention chanism to make it easy to understand and analyze. On the other hand, as a component artificial vision processing, it facilitates subsequent procedures by reducing computational s a key consideration in real-time applications. Based on the above knowledge, we propose consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, ap(SM) is obtained and the most salient region is picked up to improve detection speed. stage, a support vector machines (SVM) classifier is used to get the target quickly. 2015, xx 2 recent years, a series of simple and fast algorithm based on fourier transform was proposed, ctral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier HFT) [7]. With regards to small target detection, frequency domain method is quite different methods. It transforms the airspace information to the frequency domain, defines significant tests in the frequency domain. While, spectral residual (SR) approach does not rely on the . It calculates the difference between the original signal and a smooth one in the log amplitude nd then makes up a saliency map by transforming SR to spatial domain. PFT approach detects ts from the reconstruction that is calculated only by the phase spectrum of the input signal. e computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. ach explains the intrinsic theory of saliency detector in the frequency domain and use spectral press repeated patterns.F h numerous methods have been proposed, many of them may fail in certain circumstances, d-sky background [8], which is common in the helicopter view. In this situation, targets are d up with the background clutters in size and easily overlapped by vegetation, roads, rivers, , resulting a huge false rate in traditional algorithm. to design an appropriate method, a small target detection method inspired by the human m (HVS) has been designed in this paper. HVS is a kind of layered image processing system of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of is organized as follows. In Section 2, we describe the framework of the proposed algorithm rget detection. In Section 3, we present the experimental results. Section 4 is the conclusion le.
m Based on Human Visual System

work of the Proposed Two Stage Algorithm-A Brief Description
vides the scene into small patches and select important information through visual attention echanism to make it easy to understand and analyze. On the other hand, as a component l artificial vision processing, it facilitates subsequent procedures by reducing computational is a key consideration in real-time applications. Based on the above knowledge, we propose rk consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, map(SM) is obtained and the most salient region is picked up to improve detection speed. n stage, a support vector machines (SVM) classifier is used to get the target quickly. pkq represents the random forest predictor based on the k-th physicochemical property (cf. Equation (13)). Now, the problem is how to combine the results from the seven individual predictors to maximize the prediction quality. As indicated by a series of previous studies, using the ensemble classifier formed by fusing many individual classifiers can remarkably enhance the success rates in predicting protein subcellular localization [99,100] and protein quaternary structural attribute [101]. Encouraged by the previous investigators' studies, here we are also developing an ensemble classifier by fusing the seven individual predictors Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly.
Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.F Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. pkq pk " 1, 2,¨¨¨, 7q through a voting system, as formulated by: Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly.
Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.F Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly.

E "
Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly.
Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.F Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, a saliency map(SM) is obtained and the most salient region is picked up to improve detection speed.
In detection stage, a support vector machines (SVM) classifier is used to get the target quickly.

p1q @¨¨¨@
Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was pro such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex f transform(HFT) [7]. With regards to small target detection, frequency domain method is quite di from other methods. It transforms the airspace information to the frequency domain, defines sign target and tests in the frequency domain. While, spectral residual (SR) approach does not rely parameters. It calculates the difference between the original signal and a smooth one in the log amp spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach d small targets from the reconstruction that is calculated only by the phase spectrum of the input It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computationa HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use sp filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain circumst e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targ easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the h visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing s consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The this paper is organized as follows. In Section 2, we describe the framework of the proposed algo for small target detection. In Section 3, we present the experimental results. Section 4 is the conc of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual att selection mechanism to make it easy to understand and analyze. On the other hand, as a comp of low-level artificial vision processing, it facilitates subsequent procedures by reducing computa cost, which is a key consideration in real-time applications. Based on the above knowledge, we pr a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection a saliency map(SM) is obtained and the most salient region is picked up to improve detection In detection stage, a support vector machines (SVM) classifier is used to get the target quickly.
Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was pr such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex transform(HFT) [7]. With regards to small target detection, frequency domain method is quite d from other methods. It transforms the airspace information to the frequency domain, defines sig target and tests in the frequency domain. While, spectral residual (SR) approach does not rely parameters. It calculates the difference between the original signal and a smooth one in the log am spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach small targets from the reconstruction that is calculated only by the phase spectrum of the inpu It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computation HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use filter to suppress repeated patterns.F Although numerous methods have been proposed, many of them may fail in certain circum e.g., ground-sky background [8], which is common in the helicopter view. In this situation, tar easily mixed up with the background clutters in size and easily overlapped by vegetation, roads bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. Th this paper is organized as follows. In Section 2, we describe the framework of the proposed al for small target detection. In Section 3, we present the experimental results. Section 4 is the con of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual a selection mechanism to make it easy to understand and analyze. On the other hand, as a com of low-level artificial vision processing, it facilitates subsequent procedures by reducing compu cost, which is a key consideration in real-time applications. Based on the above knowledge, we a framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetectio a saliency map(SM) is obtained and the most salient region is picked up to improve detection In detection stage, a support vector machines (SVM) classifier is used to get the target quickly. Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier trans such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hy transform(HFT) [7]. With regards to small target detection, frequency domain meth from other methods. It transforms the airspace information to the frequency domain target and tests in the frequency domain. While, spectral residual (SR) approach d parameters. It calculates the difference between the original signal and a smooth one spectrum, and then makes up a saliency map by transforming SR to spatial domain. PF small targets from the reconstruction that is calculated only by the phase spectrum It omits the computation of SR in the amplitude spectrum, which saves about 1/3 HFT approach explains the intrinsic theory of saliency detector in the frequency dom filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in cer e.g., ground-sky background [8], which is common in the helicopter view. In this s easily mixed up with the background clutters in size and easily overlapped by veget bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method insp visual system (HVS) has been designed in this paper. HVS is a kind of layered image consisting of optical system, retina and visual pathways, which is nonuniform and no this paper is organized as follows. In Section 2, we describe the framework of the for small target detection. In Section 3, we present the experimental results. Section of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information throu selection mechanism to make it easy to understand and analyze. On the other han of low-level artificial vision processing, it facilitates subsequent procedures by redu cost, which is a key consideration in real-time applications. Based on the above know a framework consisting of two stages inspired by HVS as follows (See Figure 1). In a saliency map(SM) is obtained and the most salient region is picked up to impro In detection stage, a support vector machines (SVM) classifier is used to get the targ Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier tran such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], h transform(HFT) [7]. With regards to small target detection, frequency domain me from other methods. It transforms the airspace information to the frequency doma target and tests in the frequency domain. While, spectral residual (SR) approach parameters. It calculates the difference between the original signal and a smooth on spectrum, and then makes up a saliency map by transforming SR to spatial domain. small targets from the reconstruction that is calculated only by the phase spectrum It omits the computation of SR in the amplitude spectrum, which saves about 1/3 HFT approach explains the intrinsic theory of saliency detector in the frequency do filter to suppress repeated patterns.F Although numerous methods have been proposed, many of them may fail in c e.g., ground-sky background [8], which is common in the helicopter view. In this easily mixed up with the background clutters in size and easily overlapped by veg bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method in visual system (HVS) has been designed in this paper. HVS is a kind of layered ima consisting of optical system, retina and visual pathways, which is nonuniform and this paper is organized as follows. In Section 2, we describe the framework of th for small target detection. In Section 3, we present the experimental results. Sectio of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information thr selection mechanism to make it easy to understand and analyze. On the other h of low-level artificial vision processing, it facilitates subsequent procedures by red cost, which is a key consideration in real-time applications. Based on the above kn a framework consisting of two stages inspired by HVS as follows (See Figure 1). a saliency map(SM) is obtained and the most salient region is picked up to imp In detection stage, a support vector machines (SVM) classifier is used to get the ta pkq (17) where 015, xx 2 cent years, a series of simple and fast algorithm based on fourier transform was proposed, ral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier T) [7]. With regards to small target detection, frequency domain method is quite different ethods. It transforms the airspace information to the frequency domain, defines significant ts in the frequency domain. While, spectral residual (SR) approach does not rely on the t calculates the difference between the original signal and a smooth one in the log amplitude then makes up a saliency map by transforming SR to spatial domain. PFT approach detects from the reconstruction that is calculated only by the phase spectrum of the input signal. omputation of SR in the amplitude spectrum, which saves about 1/3 computational cost. h explains the intrinsic theory of saliency detector in the frequency domain and use spectral ess repeated patterns.R numerous methods have been proposed, many of them may fail in certain circumstances, sky background [8], which is common in the helicopter view. In this situation, targets are up with the background clutters in size and easily overlapped by vegetation, roads, rivers, esulting a huge false rate in traditional algorithm. o design an appropriate method, a small target detection method inspired by the human (HVS) has been designed in this paper. HVS is a kind of layered image processing system optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of organized as follows. In Section 2, we describe the framework of the proposed algorithm et detection. In Section 3, we present the experimental results. Section 4 is the conclusion .

Based on Human Visual System
rk of the Proposed Two Stage Algorithm-A Brief Description es the scene into small patches and select important information through visual attention chanism to make it easy to understand and analyze. On the other hand, as a component artificial vision processing, it facilitates subsequent procedures by reducing computational a key consideration in real-time applications. Based on the above knowledge, we propose consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, ap(SM) is obtained and the most salient region is picked up to improve detection speed.

2015, xx
2 recent years, a series of simple and fast algorithm based on fourier transform was proposed, ctral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier FT) [7]. With regards to small target detection, frequency domain method is quite different methods. It transforms the airspace information to the frequency domain, defines significant ests in the frequency domain. While, spectral residual (SR) approach does not rely on the It calculates the difference between the original signal and a smooth one in the log amplitude nd then makes up a saliency map by transforming SR to spatial domain. PFT approach detects ts from the reconstruction that is calculated only by the phase spectrum of the input signal. computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. ch explains the intrinsic theory of saliency detector in the frequency domain and use spectral press repeated patterns.F h numerous methods have been proposed, many of them may fail in certain circumstances, -sky background [8], which is common in the helicopter view. In this situation, targets are d up with the background clutters in size and easily overlapped by vegetation, roads, rivers, resulting a huge false rate in traditional algorithm. to design an appropriate method, a small target detection method inspired by the human m (HVS) has been designed in this paper. HVS is a kind of layered image processing system f optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of s organized as follows. In Section 2, we describe the framework of the proposed algorithm rget detection. In Section 3, we present the experimental results. Section 4 is the conclusion le.
m Based on Human Visual System ork of the Proposed Two Stage Algorithm-A Brief Description ides the scene into small patches and select important information through visual attention echanism to make it easy to understand and analyze. On the other hand, as a component l artificial vision processing, it facilitates subsequent procedures by reducing computational is a key consideration in real-time applications. Based on the above knowledge, we propose k consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, map(SM) is obtained and the most salient region is picked up to improve detection speed.
E stands for the ensemble classifier, and the symbol @ for the fusing operator. For the detailed procedures of how to fuse the results from the seven individual predictors to reach a final outcome via the voting system, see Equations (30)- (35) in [27], where a crystal clear and elegant derivation was elaborated and hence there is no need to repeat here. To provide an intuitive picture, a flowchart is given in Figure 3 to illustrate how the seven individual RF predictors are fused into the ensemble classifier. clutters. In recent years, a series of simple and fast algorithm based on fourier transform w such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hyperco transform(HFT) [7]. With regards to small target detection, frequency domain method is from other methods. It transforms the airspace information to the frequency domain, defin target and tests in the frequency domain. While, spectral residual (SR) approach does n parameters. It calculates the difference between the original signal and a smooth one in the spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT ap small targets from the reconstruction that is calculated only by the phase spectrum of the It omits the computation of SR in the amplitude spectrum, which saves about 1/3 comp HFT approach explains the intrinsic theory of saliency detector in the frequency domain an filter to suppress repeated patterns.R Although numerous methods have been proposed, many of them may fail in certain c e.g., ground-sky background [8], which is common in the helicopter view. In this situati easily mixed up with the background clutters in size and easily overlapped by vegetation, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hyperc transform(HFT) [7]. With regards to small target detection, frequency domain method i from other methods. It transforms the airspace information to the frequency domain, de target and tests in the frequency domain. While, spectral residual (SR) approach does parameters. It calculates the difference between the original signal and a smooth one in th spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT a small targets from the reconstruction that is calculated only by the phase spectrum of t It omits the computation of SR in the amplitude spectrum, which saves about 1/3 com HFT approach explains the intrinsic theory of saliency detector in the frequency domain filter to suppress repeated patterns.F Although numerous methods have been proposed, many of them may fail in certain e.g., ground-sky background [8], which is common in the helicopter view. In this situa easily mixed up with the background clutters in size and easily overlapped by vegetatio bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired (1) , D(2) means the decision made by 5, xx 2 ent years, a series of simple and fast algorithm based on fourier transform was proposed, l residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier ) [7]. With regards to small target detection, frequency domain method is quite different thods. It transforms the airspace information to the frequency domain, defines significant s in the frequency domain. While, spectral residual (SR) approach does not rely on the calculates the difference between the original signal and a smooth one in the log amplitude then makes up a saliency map by transforming SR to spatial domain. PFT approach detects rom the reconstruction that is calculated only by the phase spectrum of the input signal. mputation of SR in the amplitude spectrum, which saves about 1/3 computational cost. explains the intrinsic theory of saliency detector in the frequency domain and use spectral ss repeated patterns.R umerous methods have been proposed, many of them may fail in certain circumstances, y background [8], which is common in the helicopter view. In this situation, targets are p with the background clutters in size and easily overlapped by vegetation, roads, rivers, ulting a huge false rate in traditional algorithm.

015, xx
2 cent years, a series of simple and fast algorithm based on fourier transform was proposed, tral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier T) [7]. With regards to small target detection, frequency domain method is quite different ethods. It transforms the airspace information to the frequency domain, defines significant sts in the frequency domain. While, spectral residual (SR) approach does not rely on the t calculates the difference between the original signal and a smooth one in the log amplitude d then makes up a saliency map by transforming SR to spatial domain. PFT approach detects from the reconstruction that is calculated only by the phase spectrum of the input signal. computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. h explains the intrinsic theory of saliency detector in the frequency domain and use spectral ress repeated patterns.F numerous methods have been proposed, many of them may fail in certain circumstances, sky background [8], which is common in the helicopter view. In this situation, targets are up with the background clutters in size and easily overlapped by vegetation, roads, rivers, esulting a huge false rate in traditional algorithm.
(2), and so forth. See the text as well as Equations (11) and (16) for further explanation. The final predictor thus obtained is called "iPPBS-Opt", where "i" stands for "identify", "PPBS" for "protein-protein binding site", and "Opt" for "optimizing" training datasets. Note that the iPPBS-Opt predictor contains a parameter K, reflecting how many nearest neighbors should be considered in removing the redundant negative samples from the training dataset during the KNNC treatment (cf. Section 2.4). Its final value is determined by maximizing the overall success rate via cross-validation, as will be described later.

Result and Discussion
As pointed out in the Introduction section, one of the important procedures in developing a predictor is how to properly and objectively evaluate its anticipated success rates [12]. Towards this, we need to consider the following two aspects: one is what kind of metrics should be used to quantitatively measure the prediction accuracy; the other is what kind of test method should be adopted to derive the metrics values, as elaborated below.

Metrics for Measuring Success Rates
For measuring the success rates in identifying PPBS, a set of four metrics are usually used in literature. They are: (1) overall accuracy or Acc; (2) Mathew's correlation coefficient or MCC; (3) sensitivity or Sn; and (4) specificity or Sp (see, e.g., [102]). Unfortunately, the conventional formulations for the four metrics are not quite intuitive for most experimental scientists, particularly the one for MCC. Interestingly, by using the symbols and derivation as used in [103] for studying signal peptides, the aforementioned four metrics can be formulated by a set of equations given below [14,30,60,61,104]: ď Mcc ď 1 (18) where N`represents the total number of PPBSs investigated whereas N`the number of true PPBSs incorrectly predicted to be of non-PPBS; N´the total number of the non-PPBSs investigated whereas N´the number of non-PPBSs incorrectly predicted to be of PPBS. According to Equation (18), it is crystal clear to see the following. When N`" 0 meaning none of the true PPBSs are incorrectly predicted to be of non-PPBS, we have the sensitivity Sn " 1. When N`" N`meaning that all the PPBSs are incorrectly predicted to be of non-PPBS, we have the sensitivity Sn " 0. Likewise, when N´= 0 meaning none of the non-PPBSs are incorrectly predicted to be of PPBS, we have the specificity Sp " 1; whereas N´= N´meaning that all the non-PPBSs are incorrectly predicted to be of PPBS, we have the specificity Sp " 0. When N`" N´" 0 meaning that none of PPBSs in the positive dataset and none of the non-PPBSs in the negative dataset are incorrectly predicted, we have the overall accuracy Acc " 1 and MCC " 1; when N`" N`and N´= N´meaning that all the PPBSs in the positive dataset and all the non-PPBSs in the negative dataset are incorrectly predicted, we have the overall accuracy Acc " 0 and MCC "´1; whereas when N`" N`{2 and N´= N´{2 we have Acc " 0.5 and MCC " 0 meaning no better than random guess. As we can see from the above discussion, it would make the meanings of sensitivity, specificity, overall accuracy, and Mathew's correlation coefficient much more intuitive and easier-to-understand by using Equation (18), particularly for the meaning of MCC.
It should be pointed out, however, the set of metrics as defined in Equation (18) is valid only for the single-label systems. For the multi-label systems whose emergence has become more frequent in system biology [46,105,106] and system medicine [107], a completely different set of metrics as defined in [108] is needed.

Cross-Validation and Target Cross-Validation
Once established the evaluation metrics, the next issue is the selection of the most appropriate validation method should be used to derive the values of these metrics. Three cross-validation methods are often used to derive metrics values in statistical prediction: the independent dataset test, subsampling (or K-fold cross-validation) test, and jackknife test [109]. Of the three the jackknife test is deemed the least arbitrary as it can always yield a unique outcome for a given benchmark dataset, as elucidated in [12] and demonstrated by Equations (28)-(32) therein. Accordingly, the jackknife test has been widely recognized and increasingly used by investigators to examine the quality of various predictors (see, e.g., [46,53,54,[110][111][112][113][114][115]. However, to reduce the computational time, in this study we adopted the 10-fold cross-validation, as done by most investigators with SVM and random forests algorithms as the prediction engine. When conducting the 10-fold cross-validation for the current predictor iPPBS-Opt, however, some special consideration is needed. This is because a dataset, after optimized by the KNNC and ITHTS treatments, may miss many experimental negative samples and contain some hypothetical positive samples. It would be fine to use such a dataset to train a predictor, but not for validation. Since the validation should be conducted based on all the experimental data in the benchmark dataset but not on the added hypothetical samples nor only on the data in the reduced negative subset, a special cross-validation, the so-called target cross-validation, has been introduced here. During the target cross-validation process for the positive samples, only the experiment-confirmed samples are singled out as the targets (or test samples) for validation; but during the target cross-validation process for the negative samples, even all the excluded experimental data are taken into account. The detailed procedures of the target 10-fold cross-validation are as follows: Step 1. Before optimizing the original benchmark dataset, both its positive and negative subsets were randomly divided into 10 parts with about the same size. For example, for the all-residue benchmark dataset lgorithms 2015, xx 2 lutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, uch as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier ansform(HFT) [7]. With regards to small target detection, frequency domain method is quite different rom other methods. It transforms the airspace information to the frequency domain, defines significant rget and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the arameters. It calculates the difference between the original signal and a smooth one in the log amplitude pectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects mall targets from the reconstruction that is calculated only by the phase spectrum of the input signal. t omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. FT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral lter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, .g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are asily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, ridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human isual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system onsisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of is paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm or small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion f this article.
. Algorithm Based on Human Visual System .1. Framework of the Proposed Two Stage Algorithm-A Brief Description HVS divides the scene into small patches and select important information through visual attention election mechanism to make it easy to understand and analyze. On the other hand, as a component f low-level artificial vision processing, it facilitates subsequent procedures by reducing computational ost, which is a key consideration in real-time applications. Based on the above knowledge, we propose framework consisting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, all , after such evenly division we have: Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose all " Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose allp1q ď Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visual attention selection mechanism to make it easy to understand and analyze. On the other hand, as a component of low-level artificial vision processing, it facilitates subsequent procedures by reducing computational cost, which is a key consideration in real-time applications. Based on the above knowledge, we propose allp2q ď¨¨¨ď Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform was such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomp transform(HFT) [7]. With regards to small target detection, frequency domain method is quit from other methods. It transforms the airspace information to the frequency domain, defines target and tests in the frequency domain. While, spectral residual (SR) approach does not r parameters. It calculates the difference between the original signal and a smooth one in the log spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT appro small targets from the reconstruction that is calculated only by the phase spectrum of the in It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computat HFT approach explains the intrinsic theory of saliency detector in the frequency domain and u filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circu e.g., ground-sky background [8], which is common in the helicopter view. In this situation, easily mixed up with the background clutters in size and easily overlapped by vegetation, ro bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by visual system (HVS) has been designed in this paper. HVS is a kind of layered image process consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. this paper is organized as follows. In Section 2, we describe the framework of the proposed for small target detection. In Section 3, we present the experimental results. Section 4 is the of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through visua selection mechanism to make it easy to understand and analyze. On the other hand, as a c of low-level artificial vision processing, it facilitates subsequent procedures by reducing com cost, which is a key consideration in real-time applications. Based on the above knowledge, w allp10q " ď10 i"1 Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [ transform(HFT) [7]. With regards to small target detection, frequency domain from other methods. It transforms the airspace information to the frequency do target and tests in the frequency domain. While, spectral residual (SR) appro parameters. It calculates the difference between the original signal and a smooth spectrum, and then makes up a saliency map by transforming SR to spatial doma small targets from the reconstruction that is calculated only by the phase spec It omits the computation of SR in the amplitude spectrum, which saves abou HFT approach explains the intrinsic theory of saliency detector in the frequency filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail e.g., ground-sky background [8], which is common in the helicopter view. In easily mixed up with the background clutters in size and easily overlapped by bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection metho visual system (HVS) has been designed in this paper. HVS is a kind of layered consisting of optical system, retina and visual pathways, which is nonuniform a this paper is organized as follows. In Section 2, we describe the framework o for small target detection. In Section 3, we present the experimental results. Se of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information selection mechanism to make it easy to understand and analyze. On the oth of low-level artificial vision processing, it facilitates subsequent procedures by cost, which is a key consideration in real-time applications. Based on the abov allpiq (19) and: Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article. Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article. Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier transform such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hyperc transform(HFT) [7]. With regards to small target detection, frequency domain method is from other methods. It transforms the airspace information to the frequency domain, defi target and tests in the frequency domain. While, spectral residual (SR) approach does parameters. It calculates the difference between the original signal and a smooth one in th spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT a small targets from the reconstruction that is calculated only by the phase spectrum of t It omits the computation of SR in the amplitude spectrum, which saves about 1/3 com HFT approach explains the intrinsic theory of saliency detector in the frequency domain a filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain e.g., ground-sky background [8], which is common in the helicopter view. In this situat easily mixed up with the background clutters in size and easily overlapped by vegetatio bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired visual system (HVS) has been designed in this paper. HVS is a kind of layered image pro consisting of optical system, retina and visual pathways, which is nonuniform and nonlin this paper is organized as follows. In Section 2, we describe the framework of the prop for small target detection. In Section 3, we present the experimental results. Section 4 is of this article.  (20) where the symbol fi means that the divided 10 datasets are about the same in size, and so are their subsets.

Algorithm
Step 2. One of the 10 sets, say Algorithms 2015, xx 2 clutters. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier transform(HFT) [7]. With regards to small target detection, frequency domain method is quite different from other methods. It transforms the airspace information to the frequency domain, defines significant target and tests in the frequency domain. While, spectral residual (SR) approach does not rely on the parameters. It calculates the difference between the original signal and a smooth one in the log amplitude spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects small targets from the reconstruction that is calculated only by the phase spectrum of the input signal. It omits the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. HFT approach explains the intrinsic theory of saliency detector in the frequency domain and use spectral filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain circumstances, e.g., ground-sky background [8], which is common in the helicopter view. In this situation, targets are easily mixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method inspired by the human visual system (HVS) has been designed in this paper. HVS is a kind of layered image processing system consisting of optical system, retina and visual pathways, which is nonuniform and nonlinear. The rest of this paper is organized as follows. In Section 2, we describe the framework of the proposed algorithm for small target detection. In Section 3, we present the experimental results. Section 4 is the conclusion of this article.

Algorithm Based on Human Visual System
allp1q , was singled out as the testing dataset and the remaining nine sets as the training dataset.
Step 3. The training set was optimized using the KNNC and IHTS treatments as described in Section 2.4. After such a process, the original imbalanced training dataset would become a balanced one; i.e., its positive subset and negative subset would contain a same number of samples. Note that although the starting value for K in the KNNC treatment could be arbitrary, the following empirical approach might be of help to reduce the time for finally finding its optimal value. Suppose the starting value for K is K p0q, then we have according to our experience where N`and N´are the numbers of the total positive and negative samples in the benchmark dataset, respectively, and Int is the "integer truncation operator" meaning to take the integer part for the number in the brackets right after it [116]. Substituting the data of Equation (15) into Equation (21), we obtained K p0q " 3 or 8 for the surface-residue case or of all-residue case, respectively.
Step 4. Use the aforementioned balanced dataset to train the operation engine, followed by applying the iPPBS-Opt predictor to calculate the prediction scores for the testing dataset, which had been singled out in Step 2 before the optimized treatment and hence contained the experiment-confirmed samples only.
Step 5. Repeat Steps 2-4 until all the 10 divided sets had been singled out one-by-one for testing validation.
Step 6. Substituting the scores obtained from the above 10-round tests into Equation (18) to calculate Sn, Sp, Acc, and MCC. The metrics values thus obtained should be a function of K; for instance, the overall accuracy Acc can be expressed as Acc(K).
It is instructive to emphasize again that it is absolutely fair to use the above 10-fold cross-validation steps to compare the current predictor with the existing ones. This is because all the predictors concerned were tested using exactly the same experiment-confirmed samples and that all the added hypothetical samples had been completely excluded from the testing datasets.  (4)); and (b) the all-residue benchmark dataset (cf. Equation (5)). It can be seen from panel (a) that the overall accuracy reaches its peak at K " 9, and from panel (b) that the overall accuracy reaches its peak at K " 15 . Table 3. Comparison of the iPPBS-Opt with the other existing methods via the 10-fold cross-validation on the surface-residue benchmark dataset (Equation (4)) and the all-residue benchmark dataset (Equation (5)).  [11]; c Results obtained on the same testing dataset by the current predictor iPPBS-Opt with its parameter K " 9 for the surface-residue benchmark dataset ms 2015, xx 2

Benchmark Dataset
. In recent years, a series of simple and fast algorithm based on fourier transform was proposed, spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6], hypercomplex fourier m(HFT) [7]. With regards to small target detection, frequency domain method is quite different her methods. It transforms the airspace information to the frequency domain, defines significant nd tests in the frequency domain. While, spectral residual (SR) approach does not rely on the ters. It calculates the difference between the original signal and a smooth one in the log amplitude , and then makes up a saliency map by transforming SR to spatial domain. PFT approach detects rgets from the reconstruction that is calculated only by the phase spectrum of the input signal. the computation of SR in the amplitude spectrum, which saves about 1/3 computational cost. proach explains the intrinsic theory of saliency detector in the frequency domain and use spectral suppress repeated patterns.S ough numerous methods have been proposed, many of them may fail in certain circumstances, und-sky background [8], which is common in the helicopter view. In this situation, targets are ixed up with the background clutters in size and easily overlapped by vegetation, roads, rivers, [9], resulting a huge false rate in traditional algorithm. der to design an appropriate method, a small target detection method inspired by the human surf (cf. Equation (4)) and K " 15 for the all-residue benchmark dataset Algorithms 2015, xx clutters. In recent years, a series of simple and fast algorithm based on fourier such as spectral residual(SR) [5], phase spectrum of fourier transform(PFT) [6 transform(HFT) [7]. With regards to small target detection, frequency domain from other methods. It transforms the airspace information to the frequency do target and tests in the frequency domain. While, spectral residual (SR) appro parameters. It calculates the difference between the original signal and a smooth spectrum, and then makes up a saliency map by transforming SR to spatial doma small targets from the reconstruction that is calculated only by the phase spec It omits the computation of SR in the amplitude spectrum, which saves about HFT approach explains the intrinsic theory of saliency detector in the frequency filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail i e.g., ground-sky background [8], which is common in the helicopter view. In easily mixed up with the background clutters in size and easily overlapped by bridges [9], resulting a huge false rate in traditional algorithm.
In order to design an appropriate method, a small target detection method all (cf. Equation (5)). Also see Figure 4 for the details. Table 3 are the values of the four metrics (cf. Equation (18)) obtained by the current iPPBS-Opt predictor using the target 10-fold cross-validation on the surface-residue benchmark dataset ]. With regards to small target detection, frequency domain method is quite different . It transforms the airspace information to the frequency domain, defines significant the frequency domain. While, spectral residual (SR) approach does not rely on the lates the difference between the original signal and a smooth one in the log amplitude makes up a saliency map by transforming SR to spatial domain. PFT approach detects the reconstruction that is calculated only by the phase spectrum of the input signal. tation of SR in the amplitude spectrum, which saves about 1/3 computational cost. ains the intrinsic theory of saliency detector in the frequency domain and use spectral peated patterns.S ous methods have been proposed, many of them may fail in certain circumstances, ckground [8], which is common in the helicopter view. In this situation, targets are th the background clutters in size and easily overlapped by vegetation, roads, rivers, g a huge false rate in traditional algorithm. gn an appropriate method, a small target detection method inspired by the human ) has been designed in this paper. HVS is a kind of layered image processing system l system, retina and visual pathways, which is nonuniform and nonlinear. The rest of zed as follows. In Section 2, we describe the framework of the proposed algorithm ection. In Section 3, we present the experimental results. Section 4 is the conclusion d on Human Visual System the Proposed Two Stage Algorithm-A Brief Description scene into small patches and select important information through visual attention m to make it easy to understand and analyze. On the other hand, as a component al vision processing, it facilitates subsequent procedures by reducing computational consideration in real-time applications. Based on the above knowledge, we propose ting of two stages inspired by HVS as follows (See Figure 1). In predetection stage, ) is obtained and the most salient region is picked up to improve detection speed. a support vector machines (SVM) classifier is used to get the target quickly. ayers structure, the algorithm computational complexity becomes the prime concern. nd high processing speed, saliency detection methods in the frequency domain are surf (Equation (4)) and the all-residue benchmark dataset transform(HFT) [7]. With regards to small target detection, frequency domain method i from other methods. It transforms the airspace information to the frequency domain, de target and tests in the frequency domain. While, spectral residual (SR) approach does parameters. It calculates the difference between the original signal and a smooth one in th spectrum, and then makes up a saliency map by transforming SR to spatial domain. PFT a small targets from the reconstruction that is calculated only by the phase spectrum of t It omits the computation of SR in the amplitude spectrum, which saves about 1/3 com HFT approach explains the intrinsic theory of saliency detector in the frequency domain filter to suppress repeated patterns.S Although numerous methods have been proposed, many of them may fail in certain e.g., ground-sky background [8], which is common in the helicopter view. In this situa easily mixed up with the background clutters in size and easily overlapped by vegetatio bridges [9], resulting a huge false rate in traditional algorithm.

Listed in
In order to design an appropriate method, a small target detection method inspired visual system (HVS) has been designed in this paper. HVS is a kind of layered image pr consisting of optical system, retina and visual pathways, which is nonuniform and nonlin this paper is organized as follows. In Section 2, we describe the framework of the prop for small target detection. In Section 3, we present the experimental results. Section 4 i of this article.

Framework of the Proposed Two Stage Algorithm-A Brief Description
HVS divides the scene into small patches and select important information through selection mechanism to make it easy to understand and analyze. On the other hand, of low-level artificial vision processing, it facilitates subsequent procedures by reducing cost, which is a key consideration in real-time applications. Based on the above knowled a framework consisting of two stages inspired by HVS as follows (See Figure 1). In pre a saliency map(SM) is obtained and the most salient region is picked up to improve d In detection stage, a support vector machines (SVM) classifier is used to get the target q Due to the two layers structure, the algorithm computational complexity becomes the For the simplicity and high processing speed, saliency detection methods in the freque all (Equation (5)), respectively. See S1 Dataset for the details of the two benchmark datasets. For facilitating comparison, the corresponding results obtained by the existing methods [10,11] are also given there.
As we can see from the table, the new predictor iPPBS-Opt proposed in this paper remarkably outperformed its counterparts, particularly in Acc and MCC; the former stands for the overall accuracy, and the latter for the stability. At the first glance, although the value of Sn by Deng et al.'s method [10] is higher than that of the current predictor when tested by the surface-residue benchmark dataset, its corresponding Sp value is more than 30% lower than that of the latter, indicating the method [10] is very unstable with extremely high noise.
Because graphic approaches can provide useful intuitive insights (see, e.g., [117][118][119][120][121][122]), here we also provide a graphic comparison of the current predictor with their counterparts via the Receiver Operating Characteristic (ROC) plot [123], as shown in ( Figure 5). According to ROC [123], the larger the area under the curve (AUC), the better the corresponding predictor is. As we can see from the figure, the area under the ROC curve of the new predictor is remarkably greater than those of their counterparts fully consistent with the AUC values listed on Table 3, once again indicating a clear improvement of the new predictor in comparison with the existing ones. All the above facts have shown that iPPBS-Opt is really a very promising predictor for identifying protein-protein binding sites. Or at the very least, it can play a complementary role to the existing prediction methods in this area. Particularly, none of the existing predictors has provided a web server. In contrast to this, a user-friendly and publically accessible web server has been established for iPPBS-Opt at http://www.jci-bioinfo.cn/iPPBS-Opt, which is no doubt very useful for the majority of experimental scientist in this or related areas without the need to follow the complicated mathematical equations.
Why could the proposed method be so powerful? The reasons are as follows: First, the KNNC and IHTS treatments have been introduced to optimize the training datasets, so as to avoid many misprediction events caused by the highly imbalanced training datasets used in previous studies. Second, the ensemble technique has been utilized in this study to select the most relevant one from seven classes of different physicochemical properties. Third, the wavelets transform technique has been applied to extract some important key features, which are deeply hidden in complicated protein sequences. This is just like the studies in dealing with the extremely complicated internal motions of proteins, it is the key to grasp the low-frequency collective motion [74,75] for in-depth understanding or revealing the dynamic mechanisms of their various important biological functions [84], such as cooperative effects [78], allosteric transition [80,81], assembly of microtubules [83], and switch between active and inactive states [76]. Fourth, the PseAAC approach has been introduced to formulate the statistical samples, which has been proved very useful not only in dealing with protein/peptide sequences, but also in dealing with DNA/RNA sequences, as elaborated in a recent review paper [124].

Web Server and User Guide
To enhance the value of its practical applications, a web-server for iPPBS-Opt has been established at http://www.jci-bioinfo.cn/iPPBS-Opt. Furthermore, to maximize the convenience for the majority of experimental scientists, a step-to-step guide is provided below: Step 1. Opening the web-server at http://www.jci-bioinfo.cn/iPPBS-Opt, you will see the top page of iPPBS-Opt on your computer screen, as shown in Figure 6. Click on the Read Me button to see a brief introduction about the iPPBS-Opt predictor. Step 2. Either type or copy/paste the query protein sequences into the input box at the center of Figure 6. The input sequence should be in the FASTA format. For the examples of sequences in FASTA format, click the Example button right above the input box.
Step 3. Click on the Submit button to see the predicted result. For example, if you use the two query protein sequences in the Example window as the input, after 20 s or so, you will see the following on the screen of your computer: (1) Sequence-1 contains 109 amino acid residues, of which 11 are highlighted with red, meaning belonging to binding site; (2) Sequence-2 contains 275 residues, of which 25 are highlighted with red, belonging binding site. All these predicted results are fully consistent with experimental observations except for residues 53 in sequence-1 and residues 62 and 249 in sequence-2 that are overpredicted.
Step 4. As shown on the lower panel of Figure 6, batch prediction can also be selected by entering an e-mail address and the desired batch input file (in FASTA format naturally) via the Browse button. To see the sample of batch input file, click on the button Batch-example.
Step 5. Click on the Citation button to find the relevant papers that document the detailed development and algorithm of iPPBS-Opt.
Step 6. Click the Supporting Information button to download the benchmark dataset used in this study.

Conclusions
It is a very effective approach to optimize the training dataset via the KNNC treatment and IHTS treatment to enhance the prediction quality in identifying the protein-protein binding sites. This is because the training datasets constructed in this area without undergoing such an optimization procedure are usually extremely skewed and unbalanced, with the negative subset being overwhelmingly larger than the positive one. It is anticipated that the iPPBS-Opt web server presented in this paper will become a very useful high throughput tool for identifying protein-protein binding sites, or at the very least, a complementary tool to the existing prediction methods in this area.