Proceedings

Other

8 pages, 227 KB

Open AccessProceeding Paper

Transferable Deep Features for Keyword Spotting

by George Retsinas, Giorgos Sfikas and Basilis Gatos

Proceedings 2018, 2(2), 89; https://doi.org/10.3390/proceedings2020089 - 9 Jan 2018

Cited by 10 | Viewed by 3126

Deep features, defined as the activations of hidden layers of a neural network, have given promising results applied to various vision tasks. In this paper, we explore the usefulness and transferability of deep features, applied in the context of the problem of keyword [...] Read more.

Deep features, defined as the activations of hidden layers of a neural network, have given promising results applied to various vision tasks. In this paper, we explore the usefulness and transferability of deep features, applied in the context of the problem of keyword spotting (KWS). We use a state-of-the-art deep convolutional network to extract deep features. The optimal parameters concerning their application are subsequently studied: the impact of the choice of hidden layer, the impact of applying dimensionality reduction with a manifold learning technique, as well as the choice of dissimilarity measure used to retrieve relevant word images. Extensive numerical results show that deep features lead to state-of-the-art KWS performance, even when the test and training set come from different document collections. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

8 pages, 457 KB

Open AccessProceeding Paper

On the Joint Use of NMF and Classification for Overlapping Acoustic Event Detection

by Panagiotis Giannoulis, Gerasimos Potamianos and Petros Maragos

Proceedings 2018, 2(2), 90; https://doi.org/10.3390/proceedings2020090 - 9 Jan 2018

Viewed by 2270

Abstract

In this paper, we investigate the performance of classifier-based non-negative matrix factorization (NMF) methods for detecting overlapping acoustic events. We provide evidence that the performance of classifier-based NMF systems deteriorates significantly in overlapped scenarios in case mixed observations are unavailable during training. To [...] Read more.

In this paper, we investigate the performance of classifier-based non-negative matrix factorization (NMF) methods for detecting overlapping acoustic events. We provide evidence that the performance of classifier-based NMF systems deteriorates significantly in overlapped scenarios in case mixed observations are unavailable during training. To this end, we propose a K-means based method for artificial generation of mixed data. The method of Mixture of Local Dictionaries (MLD) is employed for the building of the NMF dictionary using both the isolated and artificially mixed data. Finally an SVM classifier is trained for each of the isolated and mixed event classes, using the corresponding MLD-NMF activations from the training set. The proposed system, tested on two experiments with (a) synthetic and (b) real events, outperforms the state-of-the-art classifier-based NMF system in the overlapped scenarios. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

8 pages, 5757 KB

Open AccessProceeding Paper

Synthetic Aperture Radar Processing for Vessel Kinematics Estimation

by Marco Reggiannini and Luigi Bedini

Proceedings 2018, 2(2), 91; https://doi.org/10.3390/proceedings2020091 - 9 Jan 2018

Cited by 7 | Viewed by 2387

Abstract

Navigating vessels leave traces of their motion in the form of wake patterns on the water surface. Wakes are visible in high resolution Synthetic Aperture Radar maps and bear information related to the kinematic variables of the vessel motion. A proper processing of [...] Read more.

Navigating vessels leave traces of their motion in the form of wake patterns on the water surface. Wakes are visible in high resolution Synthetic Aperture Radar maps and bear information related to the kinematic variables of the vessel motion. A proper processing of the wake pattern allows to estimate the vessel heading and velocity. This information can be exploited in decision procedures concerning maritime surveillance and ship traffic monitoring. This document describes the design and implementation of software procedures with the purpose of estimating the motion parameters of a navigating vessel, through processing the wake pattern generated by the vessel itself. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

5 pages, 398 KB

Open AccessProceeding Paper

Vehicle Logo Recognition with Reduced-Dimension SIFT Vectors Using Autoencoders

by Reyhan Kevser Keser, Esra Ergün and Behçet Uğur Töreyin

Proceedings 2018, 2(2), 92; https://doi.org/10.3390/proceedings2020092 - 9 Jan 2018

Cited by 5 | Viewed by 2697

Abstract

Vehicle logo recognition has become an important part of object recognition in recent years because of its usage in surveillance applications. In order to achieve a higher recognition rates, several methods are proposed, such as Scale Invariant Feature Transform (SIFT), convolutional neural networks, [...] Read more.

Vehicle logo recognition has become an important part of object recognition in recent years because of its usage in surveillance applications. In order to achieve a higher recognition rates, several methods are proposed, such as Scale Invariant Feature Transform (SIFT), convolutional neural networks, bag-of-words and their variations. A fast logo recognition method based on reduced-dimension SIFT vectors using autoencoders is proposed in this paper. Computational load is decreased by applying dimensionality reduction to SIFT feature vectors. Feature vectors of size 128 are reduced to 64 and 32 by employing two layer neural nets called vanilla autoencoders. Publicly available vehicle logo images are used for testing purposes. Results suggest that the proposed method needs half of the original SIFT based method’s memory requirement with decreased processing time per image in return of a decrease in the accuracy less than 20%. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

8 pages, 330 KB

Open AccessProceeding Paper

Sparse Representation Based Inpainting for the Restoration of Document Images Affected by Bleed-Through

by Muhammad Hanif, Anna Tonazzini, Pasquale Savino and Emanuele Salerno

Proceedings 2018, 2(2), 93; https://doi.org/10.3390/proceedings2020093 - 9 Jan 2018

Cited by 9 | Viewed by 3089

Abstract

Bleed-through is a commonly encountered degradation in ancient printed documents and manuscripts, which severely impair their readability. Digital image restoration techniques can be effective to remove or significantly reduce this degradation. In bleed-through document image restoration the main issue is to identify the [...] Read more.

Bleed-through is a commonly encountered degradation in ancient printed documents and manuscripts, which severely impair their readability. Digital image restoration techniques can be effective to remove or significantly reduce this degradation. In bleed-through document image restoration the main issue is to identify the bleed-through pixels and replace them with appropriate values, in accordance to their surroundings. In this paper, we propose a two stage method, where a pair of properly registered images of the document recto and verso is first used to locate the bleed-through pixels in each side, and then a sparse representation based image inpainting technique is used to fill-in the bleed-through areas according to the neighbourhood, in such a way to preserve the original appearance of the document. The advantages of the proposed inpainting technique over state-of-the-art methods are illustrated by the improvement in the visual results. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

6 pages, 2141 KB

Open AccessProceeding Paper

Classification of Hematoxylin and Eosin Images Using Local Binary Patterns and 1-D SIFT Algorithm

by Oğuzhan Oğuz, A. Enis Çetin and Rengul Çetin Atalay

Proceedings 2018, 2(2), 94; https://doi.org/10.3390/proceedings2020094 - 18 Jan 2018

Cited by 2 | Viewed by 2673

Abstract

In this paper, Hematoxylin and Eosin (H&E) stained liver images are classified by using both Local Binary Patterns (LBP) and one dimensional SIFT (1-D SIFT) algorithm. In order to obtain more meaningful features from the LBP histogram, a new feature vector extraction process [...] Read more.

In this paper, Hematoxylin and Eosin (H&E) stained liver images are classified by using both Local Binary Patterns (LBP) and one dimensional SIFT (1-D SIFT) algorithm. In order to obtain more meaningful features from the LBP histogram, a new feature vector extraction process is implemented for 1-D SIFT algorithm. LBP histograms are extracted with different approaches and concatenated with color histograms of the images. It is experimentally shown that,with the proposed approach, it possible to classify the H&E stained liver images with the accuracy of

88 %

. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

8 pages, 574 KB

Open AccessProceeding Paper

Towards Multimodal Surveillance for Smart Building Security

by Giuseppe Amato, Paolo Barsocchi, Fabrizio Falchi, Erina Ferro, Claudio Gennaro, Giuseppe Riccardo Leone, Davide Moroni, Ovidio Salvetti and Claudio Vairo

Proceedings 2018, 2(2), 95; https://doi.org/10.3390/proceedings2020095 - 9 Jan 2018

Cited by 7 | Viewed by 3783

Abstract

The main goal of a surveillance system is to collect information in a sensing environment and notify unexpected behavior. Information provided by single sensor and surveillance technology may not be sufficient to understand the whole context of the monitored environment. On the other [...] Read more.

The main goal of a surveillance system is to collect information in a sensing environment and notify unexpected behavior. Information provided by single sensor and surveillance technology may not be sufficient to understand the whole context of the monitored environment. On the other hand, by combining information coming from different sources, the overall performance of a surveillance system can be improved. In this paper, we present the Smart Building Suite, in which independent and different technologies are developed in order to realize a multimodal surveillance system. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

9 pages, 491 KB

Open AccessProceeding Paper

Statistical Measures: Promising Features for Time Series Based DDoS Attack Detection

by Ramin Fadaei Fouladi, Cemil Eren Kayatas and Emin Anarim

Proceedings 2018, 2(2), 96; https://doi.org/10.3390/proceedings2020096 - 10 Jan 2018

Cited by 15 | Viewed by 3139

Abstract

Data availability should be guaranteed by a web service in order to satisfy customers. One of the main challenges of information security professionals is DDoS attack which affects the availability. By masquerading itself as a legitimate user, a DDoS attacker tries to overwhelm [...] Read more.

Data availability should be guaranteed by a web service in order to satisfy customers. One of the main challenges of information security professionals is DDoS attack which affects the availability. By masquerading itself as a legitimate user, a DDoS attacker tries to overwhelm a server by sending a great number of useless packets that influences the quality of service (QoS) of the network. DDoS attack can result in a great damage to network services. Useless packets similar to normal ones are dispatched by the attacker which leaves the intrusion detection system impotent of detection. Transferring from conventional packet-based analysis methods to time series based (flow-based) algorithms would be a promising alternative to spot DDoS attacks. In this work, we extract four measures of periodicity, kurtosis, skewness and self-similarity of a time series and investigate the performance of these parameters in separating DDoS attack from normal traffic. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

7 pages, 1062 KB

Open AccessProceeding Paper

Size and Heading of SAR-Detected Ships through the Inertia Tensor

by Luigi Bedini, Marco Righi and Emanuele Salerno

Proceedings 2018, 2(2), 97; https://doi.org/10.3390/proceedings2020097 - 9 Jan 2018

Cited by 8 | Viewed by 2945

Abstract

We present a strategy to estimate the heading, the length overall and the beam overall of targets already detected as ships in a wide-swath SAR image acquired by a satellite platform. Such images are often affected by distortions due to marine clutter, spectral [...] Read more.

We present a strategy to estimate the heading, the length overall and the beam overall of targets already detected as ships in a wide-swath SAR image acquired by a satellite platform. Such images are often affected by distortions due to marine clutter, spectral leakage, or antenna sidelobes. These can mask the target image, thus hampering the possibility of evaluating the size and the behaviour of the ship. Even in the presence of strong artefacts, we found that the principal inertia axes can help the estimation of the target heading and be included in an iterative procedure to erode the false target features, so to enable a more accurate evaluation of the overall measurements of the ship. Here we introduce our idea and present some results obtained from real SAR images. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

7 pages, 1858 KB

Open AccessProceeding Paper

Environmental Monitoring Integrated with a Proactive Marine Information System

by Davide Moroni, Gabriele Pieri, Marco Tampucci and Ovidio Salvetti

Proceedings 2018, 2(2), 98; https://doi.org/10.3390/proceedings2020098 - 8 Jan 2018

Cited by 4 | Viewed by 2492

Abstract

In the framework of environmental monitoring, the remote detection and monitoring of oil spills at sea is an important ability due to the high demand of oil based products. This situation causes, that shipping routes are very crowded and the likelihood of oil [...] Read more.

In the framework of environmental monitoring, the remote detection and monitoring of oil spills at sea is an important ability due to the high demand of oil based products. This situation causes, that shipping routes are very crowded and the likelihood of oil slicks occurring is also increasing. In this paper we propose a fully integrated and inter-operable information system which can act as a valuable monitoring tool. Such a marine information system is able to monitor ship traffic and marine operators by integrating heterogeneous signals and data through sensing capabilities from a variety of electronic sensors, along with geo-positioning tools, and through a communication infrastructure. This system is able to transfer the integrated data, freely and seamlessly, between different elements of the system itself (and their users). The system also provides a set of decision support services capable of performing functionalities which can act as a support for decision makers. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

8 pages, 599 KB

Open AccessProceeding Paper

Computational Topology to Monitor Human Occupancy

by Paolo Barsocchi, Pietro Cassará, Daniela Giorgi, Davide Moroni and Maria Antonietta Pascali

Proceedings 2018, 2(2), 99; https://doi.org/10.3390/proceedings2020099 - 10 Jan 2018

Cited by 4 | Viewed by 2501

Abstract

The recent advances in sensing technologies, embedded systems, and wireless communication technologies, make it possible to develop smart systems to monitor human activities continuously. The occupancy of specific areas or rooms in a smart building is an important piece of information, to infer [...] Read more.

The recent advances in sensing technologies, embedded systems, and wireless communication technologies, make it possible to develop smart systems to monitor human activities continuously. The occupancy of specific areas or rooms in a smart building is an important piece of information, to infer the behavior of people, or to trigger an advanced surveillance module. We propose a method based on computational topology to infer the occupancy of a room monitored for a week by a system of low-cost sensors. Full article

(This article belongs to the Proceedings of International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM))

► Show Figures

Figure 1

Journal Menu

Journal Browser

Proceedings, 2018, IWCIM 2017

International Workshop on Computational Intelligence for Multimedia Understanding (IWCIM)

Other

Further Information

Guidelines

MDPI Initiatives

Follow MDPI