# Unsteady State Lightweight Iris Certification Based on Multi-Algorithm Parallel Integration

^{1}

^{2}

^{3}

^{*}

## Abstract

**:**

## 1. Introduction

## 2. Multi-Algorithm Parallel Integration Model and Algorithm Prerequisites

#### 2.1. Multi-Algorithm Parallel Integration Model Structure

#### 2.2. Algorithm Structure and Overall Idea Introduction

## 3. Lightweight Unsteady State Constrained Iris Certification

**Image processing layer:**Biometric information focuses on the uniqueness and stability of features, such as the shape and size of facial features in face recognition [30]. The variation of the amount of information in each part of the iris can be used as a unique iris feature [31], because the relative change relationship among the iris key points rarely changes greatly, regardless of the change of the external environment and the acquisition status. Therefore, the stability of the features means that the information measured or obtained is robust to certain distortions.

**Feature recognition layer:**Gabor and Haar are the most commonly used iris feature extraction filters. Their performance has been fully tested, and they have certain iris rotation invariance. Because the method of extracting the information variation law is used in this paper, the Gabor filter and Haar wavelet are more suitable for the image processed by the image processing layer. The Gabor filter performs iris certification by converting the filter direction and frequency into a binary form, which is a carrier code and is more suitable for the Hamming distance model. The Haar wavelet uses the amount of the block information as the iris feature, and the neural network is more suitable for this. The appropriate feature extraction and recognition algorithms are selected for the problem of phase deflection during iris acquisition in the feature recognition layer, and the differences between various categories of irises are improved, the features of the same iris category can be gathered, and the interference of phase deflection can also be suppressed as much as possible.

**Decision layer:**The illumination condition can affect the appearance of iris texture, so that the iris texture change law is not obvious. Although the image processing layer and the feature recognition layer can eliminate some external environmental influences, the algorithm itself is different for the external illumination environment, and the decision layer is different. The function is to set the algorithm to have the maximum number of correct certifications to be the most trusted algorithm in the illumination environment by statistically evaluating the algorithm’s recognition results under certain circumstances in a certain external environments (illumination). The certification result of the algorithm is taken as the final certification result under the illumination condition.

#### 3.1. Image Processing Layer

#### 3.2. Feature Recognition Layer

#### 3.2.1. Gabor + Hamming

_{1}to G

_{16}. Each Gabor processed image G

_{n}is equally divided into 8 × 4 sub-graphs of 32 × 8 dimensions. The sub-graphs are numbered B

_{n-1}~B

_{n-32}. The feature code is set by using the over-threshold judgment method. The determination threshold is set M

_{1}. The average amplitude value T of each sub-graph is compared with M

_{1}(the decision threshold is calculated from the training iris library that meets the prerequisites for this paper, which is the best value for most iris certifications). If the amplitude value is smaller than or equal to M

_{1}, the feature code of this sub-graph is set to 0. If the amplitude value is bigger than or equal to M

_{1}, the feature code of this sub-graph is set to 1. Finally, there is a 512-bit (32 × 16) binary feature code. The Hamming distance between the test iris and the template iris is calculated as a feature similarity. The feature similarity calculation formula of Gabor + Hamming is shown in Equation (4).

#### 3.2.2. Haar + BP

#### 3.3. Decision Layer

## 4. Experiments and Analysis

**Data acquisition:**The JLU iris library of Jilin University, China [36] and CASIA iris library of Chinese Academy of Sciences, China [37] are used in all experiments in this paper. The JLU iris library is collected by the Biometrics and Information Security Technology Laboratory of Jilin University and generated by video screenshots. As of 2019, there are more than 80 categories of irises in the original iris library, and each category has more than 1000 images of various states. The number of images of unsteady irises is still expanding. The CASIA Iris Library is a commonly used iris library in the world of iris recognition and has been released for four generations.

**Experimental external environment:**In the experiments, the CPU frequency is dual-core 2.5 GHz, the memory is 8 GB, and the operating system is Windows.

**Evaluation metrics:**The evaluation indexes are the ROC curve [38] (a curve indicating the relationship between the False Reject Rate (FRR) and the False Accept Rate (FAR)), the Equal Error Rate (EER) (the value where FRR is equal to the FAR), and Correct Recognition Rate (CRR) [39].

**Experimental setup:**The experiments first self-certified the contribution of each layer to the certification and explained the relationship between the algorithm settings of each layer and the actual prerequisites. As set out in Section 4.1, the image processing layer experiment is mainly concerned with the extraction of defocused conditions, and explores the effect of extracted iris texture on iris certification; as set out in Section 4.2, feature recognition layer experiment was mainly designed to conduct tests on iris deflection and explore its impact on iris certification. As explained in Section 4.3, the decision layer experiment aimed mainly to test for the presence or absence of illumination and explore the impact of the statistical decision-making method based on external illumination on iris certification. Finally, a comprehensive experiment is set out in Section 4.4. As set out in Section 4.4.1, the actual effect of the overall model is tested, which can explain how the algorithm structure improves the overall recognition rate in the case where the traditional algorithms are used in each layer. As set out in Section 4.4.2, the performance of the algorithm is analyzed by comparing it to existing algorithms according to the prerequisites of this paper.

#### 4.1. Image Processing Layer Experiment

#### 4.2. Feature Recognition Layer Experiment

- The iris recognition algorithm based on Zernike moment phase feature [40];
- The iris recognition algorithm based on deep learning architecture [41];
- The iris recognition algorithm based on statistically characteristic center symmetric local binary pattern (SCCS-LBP) [42];
- The secondary iris recognition algorithm based on BP neural networks [11];
- The iris recognition algorithm based on cross-spectral matching [43];
- Iris recognition based on the Gabor + Hamming algorithm in the recognition layer of this study;
- Iris recognition based on the Haar + BP algorithm in the recognition layer of this study.

#### 4.3. Decision Layer Experiment

#### 4.4. Certification Structure of Comprehensive Effect Experiment

#### 4.4.1. Meaning of Tightness Degree for Each Layer

#### 4.4.2. Existing Method Comparison Experiment

## 5. Conclusions

## Author Contributions

## Funding

## Conflicts of Interest

## References

- Minaee, S.; Abdolrashidi, A.A.; Wang, Y. Iris Recognition Using Scattering Transform and Textural Features. In Proceedings of the 2015 IEEE Signal Processing and Signal Processing Education Workshop (SP/SPE) IEEE, Salt Lake City, UT, USA, 9–12 August 2015. [Google Scholar]
- Gao, S.; Zhu, X.; Liu, Y.; He, F.; Huo, G. A quality assessment method of Iris image based on support vector machine. J. Fiber Bioeng. Informat.
**2015**, 8, 293–300. [Google Scholar] - Lee, Y.W.; Kim, K.W.; Hoang, T.M.; Arsalan, M.; Park, K.R. Deep residual CNN-based ocular recognition based on rough pupil detection in the images by NIR camera sensor. Sensors
**2019**, 19, 842. [Google Scholar] [CrossRef] [PubMed] - Bazrafkan, S.; Thavalengal, S.; Corcoran, P. An end to end deep neural network for iris segmentation in unconstrained scenarios. Neural Netw.
**2018**, 106, 79–95. [Google Scholar] [CrossRef] [PubMed] - Liu, N.; Liu, J.; Sun, Z.; Tan, T. A code-level approach to heterogeneous iris recognition. IEEE Trans. Inf. Forensics Secur.
**2017**, 12, 2373–2386. [Google Scholar] [CrossRef] - Liu, J.; Sun, Z.; Tan, T. Distance metric learning for recognizing low-resolution iris images. Neurocomputing
**2014**, 144, 484–492. [Google Scholar] [CrossRef] - Yang, D.T.; Fan, P.C. Effective Scale-Invariant Feature Transform Based Iris Matching Technology for Identity Identification. In Proceedings of the IEEE International Conference on Consumer Electronics-Taiwan (ICCE-TW), Taichung, Taiwan, 19–21 May 2018. [Google Scholar]
- Vyas, R.; Kanumuri, T.; Sheoran, G.; Dubey, P. Recent trends of ROI segmentation in iris biometrics: A survey. Int. J. Biometrics
**2019**, 11, 274. [Google Scholar] [CrossRef] - Redman, A.; Zainal, N.; Suandi, S.A. Automated segmentation of iris images acquired in an unconstrained environment using HOG-SVM and GrowCut. Digit. Signal Process.
**2019**, 11, 274–307. [Google Scholar] [CrossRef] - Han, C.Y.; Kwon, S.H.; Choi, H.I.; Lee, S.J.; Wee, N.S. Seeded Ising model and the statistical nature of human iris templates. Phys. Rev. E
**2018**, 98, 032115. [Google Scholar] [CrossRef] [Green Version] - Liu, S.; Liu, Y.N.; Zhu, X.D.; Huo, G.; Liu, W.T.; Feng, J.K. Iris double recognition based on modified evolutionary neural network. J. Electron. Imag.
**2017**, 26, 063023. [Google Scholar] [CrossRef] - Singh, G.; Deep, K. Effectiveness of new Multiple-PSO based membrane optimization algorithms on CEC 2014 benchmarks and iris classification. Nat. Comput.
**2017**, 16, 473–496. [Google Scholar] [CrossRef] - Mahesh Kumar, S.V.; Gunasundari, R. Computer-aided diagnosis of anterior segment eye abnormalities using visible wavelength image analysis based machine learning. J. Med. Syst.
**2018**, 42, 128. [Google Scholar] - Yu, Q.; Tang, H.; Tan, K.C.; Yu, H. A brain-inspired spiking neural network model with temporal encoding and learning. Neurocomputing
**2014**, 138, 3–13. [Google Scholar] [CrossRef] - Yu, H.; Zhang, X.J.; Wang, S.; Song, S.M. Alternative framework of the Gaussian filter for non-linear systems with synchronously correlated noises. IET Sci. Meas. Technol.
**2016**, 10, 306–315. [Google Scholar] [CrossRef] - Li, Z.X.; Liu, J.M.; Li, S.; Bai, D.Y. Group targets tracking algorithm based on box particle filter. Acta Autom. Sin.
**2015**, 41, 785–798. [Google Scholar] - Lee, G.S.; Lee, S.H.; Kim, G.O.; Park, J.H.; Park, Y.S. A modified GrabCut using a clustering technique to reduce image noise. Symmetry
**2016**, 8, 64. [Google Scholar] [CrossRef] - Lei, S.; Wang, Z.; Xu, R. Image enhancement for surveillance video of coal mining face based on single-scale retinex algorithm combined with bilateral filtering. Symmetry
**2017**, 9, 93. [Google Scholar] - Völgyes, D.; Martinsen, A.C.T.; Stray-Pedersen, A.; Waaler, D.; Pedersen, M. A weighted histogram-based tone mapping algorithm for CT images. Algorithms
**2018**, 11, 111. [Google Scholar] [CrossRef] - Xu, S.; Hu, L.; Yang, X. Quality-aware features-based noise level estimator for block matching and three-dimensional filtering algorithm. J. Electron. Imag.
**2016**, 25, 013029. [Google Scholar] [CrossRef] - Tajouri, I.; Aydi, W.; Ghorbel, A.; Masmoudi, N. Efficient iris texture analysis method based on Gabor ordinal measures. J. Electron. Imag.
**2017**, 26, 043012. [Google Scholar] [CrossRef] - Pruthi, M.; Pankaj. The minimum hamming distances of the irreducible cyclic codes of length. J. Discret. Math. Sci. Cryptogr.
**2016**, 19, 965–995. [Google Scholar] [CrossRef] - Lim, S.; Lee, K.; Byeon, O.; Kim, T. Efficient iris recognition through improvement of feature vector and classifier. Etri J.
**2011**, 23, 61–70. [Google Scholar] [CrossRef] - Lin, H. Feature Input Symmetry Algorithm of Multi-Modal Natural Language Library Based on BP Neural Network. Symmetry
**2019**, 11, 341. [Google Scholar] [CrossRef] - Kaur, B.; Singh, S.; Kumar, J. Robust Iris Recognition Using Moment Invariants. Wireless Pers. Comm.
**2018**, 99, 799–828. [Google Scholar] [CrossRef] - Liu, S.; Liu, Y.; Zhu, X.; Liu, W.; Cheng, Y. Sequence Iris quality evaluation algorithm based on morphology and gray distribution. J. Jilin Univ.
**2018**, 56, 1156–1162. [Google Scholar] - Yu, L.; Zhang, X.; Liu, Y.; Zhu, X. Iris location algorithm based on small regional search. J. Comput. Inform. Syst.
**2015**, 11, 4221–4228. [Google Scholar] - Karakaya, M.; Celik, E.T. Effect of pupil dilation on off-angle iris recognition. J. Electron. Imag.
**2019**, 28, 033022. [Google Scholar] [CrossRef] - Liu, S.; Liu, Y.; Zhu, X.; Feng, J.; Lu, S. Iris location algorithm based on partitioning search. Comput. Eng. Appl.
**2018**, 54, 212–217. [Google Scholar] - Tome, P.; Vera-Rodriguez, R.; Fierrez, J.; Ortega-Garcia, J. Facial soft biometric features for forensic face recognition. Forensic Sci. Int.
**2015**, 257, 271–284. [Google Scholar] [CrossRef] [Green Version] - Liu, S.; Liu, Y.; Zhu, X.; Liu, Z.; Huo, G.; Ding, T.; Zhang, K. Gabor Filtering and Adaptive Optimization Neural Network for Iris Double Recognition. In Proceedings of the 13th Chinese Conference on Biometric Recognition (CCBR 2018), Urumqi, China, 11–12 August 2018; pp. 441–449. [Google Scholar]
- Liu, Y.; Liu, S.; Zhu, X.; Chen, Y.; Zheng, S.; Shen, C. LOG operator and adaptive optimization Gabor filtering for iris recognition. J. Jilin Univ.
**2018**, 48, 1606–1613. [Google Scholar] - Bouzenad, K.; Ramdani, M. Multivariate statistical process control using enhanced bottleneck neural network. Algorithms
**2017**, 10, 49. [Google Scholar] [CrossRef] - Shao, X.; Wang, H.; Liu, J.; Tang, J.; Li, J.; Zhang, X.; Shen, C. Sigmoid function based integral-derivative observer and application to autopilot design. Mech. Syst. Signal Process.
**2017**, 84, 113–127. [Google Scholar] [CrossRef] - Yuan, C.; Sun, X.; Wu, Q.M.J. Difference co-occurrence matrix using BP neural network for fingerprint liveness detection. In Soft Computing; Springer: Berlin, Germany, 2019; Volume 23, pp. 5157–5169. [Google Scholar]
- JLU Iris Image Database. Available online: http://www.jlucomputer.com/index/irislibrary/irislibrary.html (accessed on 11 September 2018).
- CASIA Iris Image Database. Available online: http://www.cbsr.ia.ac.cn/china/Iris%20Databases%20CH. asp (accessed on 11 September 2018).
- Taher, F.; Werghi, N.; Al-Ahmad, H. Computer aided diagnosis system for early lung cancer detection. Algorithms
**2015**, 8, 1088–1110. [Google Scholar] [CrossRef] - Liu, S.; Liu, Y.; Zhu, X.; Lin, Z.; Yang, J. Ant colony mutation particle swarm optimization for secondary iris recognition. J. Comput. -Aided Des. Comput. Graph.
**2018**, 30, 1604–1614. [Google Scholar] [CrossRef] - Tan, C.W.; Kumar, A. Accurate iris recognition at a distance using stabilized iris encoding and zernike moments phase features. IEEE Trans. Image Process.
**2014**, 23, 3962–3974. [Google Scholar] [CrossRef] [PubMed] - He, F.; Han, Y.; Wang, H.; Ji, J.; Liu, Y.; Ma, Z. Deep learning architecture for iris recognition based on optimal Gabor filters and deep belief network. J. Electron. Imag.
**2017**, 26, 023005. [Google Scholar] [CrossRef] - Li, H.L.; Guo, L.H.; Li, X.M.; Wang, X.Z. Iris recognition based on SCCS-LBP. Optic. Precis. Eng.
**2013**, 21, 2129–2136. [Google Scholar] - Nalla, P.R.; Kumar, A. Toward more accurate iris recognition using cross-spectral matching. IEEE Trans. Image Process.
**2016**, 26, 208–221. [Google Scholar] [CrossRef] [PubMed] - Pineiro-Chousa, J.; Vizcaino-Gonzalez, M.; Ribeiro-Navarrete, S. Using voting decisions to identify shocks in the financial services industry. In Service Business; Springer: Berlin, Germany, 2019; Volume 13, pp. 419–431. [Google Scholar]
- Viriri, S.; Tapamo, J. Iris pattern recognition based on cumulative sums and majority vote methods. Int. J. Adv. Robot. Syst.
**2017**, 14, 1–9. [Google Scholar] [CrossRef] - Das, A.; Pal, U.; Ferrer, M.A.; Blumenstein, M. A Decision-level Fusion Strategy for Multimodal Ocular Biometric in Visible Spectrum Based on Posterior Probability. In Proceedings of the IEEE International Joint Conference on Biometrics (IJCB), Denver, CO, USA, 1–4 October 2017. [Google Scholar]
- Minaee, S.; Abdolrashidiy, A.; Wang, Y. An Experimental Study of Deep Convolutional Features for Iris Recognition. In Proceedings of the IEEE Signal Processing in Medicine and Biology Symposium (SPMB), Philadelphia, PA, USA, 3 December 2016. [Google Scholar]
- Ma, T.; Xiao, F. An improved method to transform triangular fuzzy number into basic belief assignment in evidence theory. IEEE Access
**2019**, 7, 25308–25322. [Google Scholar] [CrossRef] - Menon, H.; Mukherjee, A. Iris Biometrics Using Deep Convolutional Networks. In Proceedings of the IEEE International Instrumentation and Measurement Technology Conference (I2MTC), Houston, TX, USA, 14–17 May 2018. [Google Scholar]
- Liu, Y.; Liu, S.; Zhu, X.; Huo, G.; Ding, T.; Zhang, K.; Jiang, X.; Guo, S.; Zhang, Q. Iris secondary recognition based on decision particle swarm optimization and stable texture. J. Jilin Univ.
**2019**, 49, 1329–1338. [Google Scholar] - Liu, Y.; Liu, S.; Zhu, X.; Liu, T.; Yang, X. Iris recognition algorithm based on feature weighted fusion. J. Jilin Univ.
**2019**, 49, 221–229. (In Chinese) [Google Scholar] - Zhang, S.; Poon, S.K.; Vuong, K.; Sneddon, A.; Loy, C.T. A deep learning-based approach for gait analysis in huntington disease. Stud. Health Technol. Inform.
**2019**, 264, 477–481. [Google Scholar] [PubMed] - Arcos-Garcia, A.; Alvarez-Garcia, J.A.; Soria-Morillo, L.M. Evaluation of deep neural networks for traffic sign detection systems. Neurocomputing
**2018**, 316, 332–344. [Google Scholar] [CrossRef] - Dang, L.; Hassan, S.; Im, S.; Lee, J.; Lee, S.; Moon, H. Deep learning based computer generated face identification using convolutional neural network. Appl. Sci.
**2018**, 8, 2610. [Google Scholar] [CrossRef] - Hong, D.; Liu, W.; Wu, X.; Pan, Z.; Su, J. Robust palmprint recognition based on the fast variation Vese-Osher Model. Neurocomputing
**2016**, 174, 999–1012. [Google Scholar] [CrossRef] - Li, J.; Feng, J.; Kuo, C.-C.J. Deep convolutional neural network for latent fingerprint enhancement. Signal Process Image Comm.
**2018**, 60, 52–63. [Google Scholar] [CrossRef] - Long, M.; Zeng, Y. Detecting iris liveness with batch normalized convolutional neural network. CMC-Comput. Mater. Con.
**2019**, 58, 493–504. [Google Scholar] [CrossRef]

**Figure 1.**Results of each step of the iris-processing stage: (

**a**) quality qualified image; (

**b**) iris area (two inner regions of the white ring); (

**c**) normalized image; (

**d**) enhanced image; (

**e**) recognition area.

**Figure 3.**Examples of eye images of the same person at different shooting times: (

**a**,

**b**) clear grayscale image with normal light conditions; (

**c**,

**d**) clear grayscale image with darker and brighter light conditions, respectively; (

**e**,

**f**) blurred grayscale images.

**Figure 5.**Unsteady change of constrained iris texture taken by the same person: (

**a**) normal; (

**b**) normal illumination clear and undeflected; (

**c**) normal illumination blur and deflected; (

**d**) dark illumination clear and undeflected.

**Figure 7.**Original image and six different filtered processed images: (

**a**) Iris original image; (

**b**) Gaussian filter; (

**c**) Square filter; (

**d**) Median filter; (

**e**) Bilateral filter; (

**f**) Equalization histogram; (

**g**) Laplacian of Gaussian.

**Figure 8.**Six feature difference images: (

**a**) Gaussian filter; (

**b**) Square filter; (

**c**) Median filter; (

**d**) Bilateral filter; (

**e**) Equalization histogram; (

**f**) Laplacian of Gaussian.

**Figure 9.**Stable feature recognition area of the same category and different category: (

**a**) same category 1; (

**b**) same category 2; (

**c**) different category.

**Figure 10.**Haar wavelet and the block of each sub-blocks: (

**a**) Haar wavelet; (

**b**) Sub-block distribution of G1; (

**c**) Sub-block distribution of G2; (

**d**) Sub-block distribution of G3.

Category Number | Iris Images per Category | Total Number of Images | Matches within Category | Matches outside Category | Total Number of Matches |
---|---|---|---|---|---|

100 | 10 | 1000 | 1600 | 5130 | 6730 |

No Processing | Gaussian Filtering | Log Operator and Equalization Histogram | Ensemble Model | ||||
---|---|---|---|---|---|---|---|

CRR | EER | CRR | EER | CRR | EER | CRR | EER |

97.45% | 2.88% | 98.76% | 1.43% | 99.03% | 0.89% | 99.41% | 0.48% |

Category Number | Iris Images in per Category | Total Number of Images | Matches within Category | Matches outside Category | Total Number of Matches |
---|---|---|---|---|---|

50 | 100 | 5000 | 8765 | 20623 | 29388 |

Zernike | Deep Learning Architecture | SCCS-LBP | Gabor + Hamming | ||||

CRR | EER | CRR | EER | CRR | EER | CRR | EER |

98.73% | 1.47% | 96.78% | 3.58% | 97.43% | 2.63% | 97.79% | 2.31% |

Secondary Iris Recognition | Cross-Spectral Matching | Ensemble Model | Haar + BP | ||||

CRR | EER | CRR | EER | CRR | EER | CRR | EER |

99.14% | 0.84% | 97.98% | 2.05% | 99.48% | 0.57% | 98.18% | 1.79% |

Category Number | Iris Images in per Category | Total Number of Images | Matches within Category | Matches outside Category | Total Number of Matches |
---|---|---|---|---|---|

100 | 20 | 2000 | 2645 | 8156 | 10801 |

Decision Voting | Posterior Probability Decision | Cumulative Sums and Majority Vote | Ensemble Model | ||||
---|---|---|---|---|---|---|---|

CRR | EER | CRR | EER | CRR | EER | CRR | EER |

98.42% | 1.81% | 97.83% | 2.64% | 95.76% | 4.54% | 99.18% | 0.72% |

Iris Library | Category Number | Iris Images in per Category | Total Number of Images | Matches within Category | Matches Outside Category | Total Number of Matches |
---|---|---|---|---|---|---|

CASIA-V1 | 120 | 5 | 600 | 1200 | 6000 | 7200 |

Iris-Interval | 200 | 5 | 1000 | 2000 | 5460 | 7460 |

Iris-Lamp | 400 | 5 | 2000 | 3500 | 6500 | 10000 |

CASIA-V1 | CASIA-Iris-Interval | CASIA-Iris-Lamp | |||
---|---|---|---|---|---|

CRR | EER | CRR | EER | CRR | EER |

99.78% | 0.38% | 99.37% | 0.52% | 99.13% | 0.85% |

Iris library | CASIA-V1 | CASIA-Iris-Interval | CASIA-Iris-Lamp |
---|---|---|---|

Deep learning neural network [47] | 76.68% | 74.69% | 79.84% |

Evidence theory certification by clustering method [48] | 86.45% | 83.78% | 81.76% |

Traditional convolutional neural network [49] | 83.12% | 85.46% | 78.42% |

Decision particle swarm optimization algorithm and stable feature [50] | 91.23% | 90.75% | 88.73% |

Multi-feature weighted fusion [51] | 69.75% | 70.86% | 73.42% |

VGG16 [52] | 80.74% | 83.47% | 84.23% |

Faster R-CNN Inception Resnet V2 [53] | 93.42% | 87.65% | 94.31% |

Algorithm in this paper | 99.78% | 99.37% | 99.13% |

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Shuai, L.; Yuanning, L.; Xiaodong, Z.; Kuo, Z.; Tong, D.; Xinlong, L.; Chaoqun, W.
Unsteady State Lightweight Iris Certification Based on Multi-Algorithm Parallel Integration. *Algorithms* **2019**, *12*, 194.
https://doi.org/10.3390/a12090194

**AMA Style**

Shuai L, Yuanning L, Xiaodong Z, Kuo Z, Tong D, Xinlong L, Chaoqun W.
Unsteady State Lightweight Iris Certification Based on Multi-Algorithm Parallel Integration. *Algorithms*. 2019; 12(9):194.
https://doi.org/10.3390/a12090194

**Chicago/Turabian Style**

Shuai, Liu, Liu Yuanning, Zhu Xiaodong, Zhang Kuo, Ding Tong, Li Xinlong, and Wang Chaoqun.
2019. "Unsteady State Lightweight Iris Certification Based on Multi-Algorithm Parallel Integration" *Algorithms* 12, no. 9: 194.
https://doi.org/10.3390/a12090194