IrisCodeNet：虹膜特征编码网络

doi:10.3778/j.issn.1002-8331.2011-0466

摘要/Abstract

摘要： 使用有效的特征提取算法对虹膜纹理进行准确的表达是虹膜识别技术的关键。基于虹膜识别任务的特殊性，提出了用于虹膜特征编码的网络模型IrisCodeNet。该网络架构使用了改进的BasicBlock，并结合了可以扩大决策边界的损失函数AM-Softmax（additive margin softmax）。为了获取最佳的虹膜识别效果，对AM-Softmax的参数设置、虹膜图像预处理输入形式、数据增强方式、网络输入尺寸做了细致的研究。实验结果表明：使用IrisCodeNet训练得到的特征提取器在CASIA-Iris-Thousand、CASIA-Iris-Distance、IITD虹膜数据库上进行测试，所评估的等错误率（equal error rate，EER）和正确接受率（true acceptance rate，TAR）均远远超过了广泛应用的传统算法。特别地，IrisCodeNet无需传统的虹膜归一化或精确的虹膜分割步骤依然取得了极好的识别效果。并且使用Grad-CAM（gradient-weighted class activation mapping）算法进行了可视化分析，结果表明该网络框架有效地关注了虹膜纹理信息，从而证明了IrisCodeNet具有较强的虹膜纹理特征提取能力。

关键词: 虹膜识别, 特征编码, 图像预处理, AM-Softmax, Grad-CAM

Abstract: Using effective feature extraction algorithm to accurately represent the texture of iris is the key to iris recognition technology. Based on the particularity of the task of iris recognition, IrisCodeNet, a network model for iris feature coding, is presented in this paper. The network architecture uses an improved BasicBlock combined with a loss function AM-Softmax（additive margin softmax） that can expand the decision boundary. What’s more, in order to get the best iris recognition performance, the parameters settings of AM-Softmax, the input forms of iris image preprocessed, the enhancement methods of data and the input sizes of network are studied in detail. The experimental results show that the feature extractor trained by IrisCodeNet is tested on CASIA-Iris-Thousand, CASIA-Iris-Distance and IITD iris databases. The EER（equal error rate） and TAR（true acceptance rate） evaluated are far superior to the used widely traditional algorithms. In particular, IrisCodeNet still achieves excellent recognition results without the steps of traditional iris normalization or precise iris segmentation. The visualization analysis using Grad-CAM（gradient-weighted class activation mapping） algorithm shows that the network framework pays attention to the texture of iris effectively, which proves that IrisCodeNet has a strong ability to extract features of iris.

Key words: iris recognition, feature coding, image preprocessing, AM-Softmax, Grad-CAM

贾丁丁, 沈文忠. IrisCodeNet：虹膜特征编码网络[J]. 计算机工程与应用, 2022, 58(10): 185-192.

JIA Dingding, SEHN Wenzhong. IrisCodeNet：Iris Feature Coding Network[J]. Computer Engineering and Applications, 2022, 58(10): 185-192.

参考文献

[1] DAUGMAN J.How iris recognition works[J].IEEE Transactions on Circuits and Systems for Video Technology，2004，14（1）：21-30.
[2] MASEK L.Recognition of human iris patterns for biometric identification[D].Perth：The University of Western Australia，2003.
[3] MONRO D M，RAKSHIT S，ZHANG D.DCT-based iris recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2007，29（4）：586-595.
[4] MIYAZAWA K，ITO K，AOKI T，et al.An efficient iris recognition algorithm using phase-based image matching[C]//2005 IEEE International Conference on Image Processing，2005：49-52.
[5] SUN Z，TAN T.Ordinal measures for iris recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2009，31（12）：2211-2226.
[6] BOYD A，CZAJKA A，BOWYER K.Are Gabor kernels optimal for iris recognition?[C]//2020 IEEE International Joint Conference on Biometrics，2020：1-9.
[7] GANGWAR A，JOSHI A.DeepIrisNet：deep iris representation with applications in iris recognition and cross-sensor iris recognition[C]//2016 IEEE International Conference on Image Processing，2016.
[8] GANGWAR A，JOSHI A，JOSHI P，et al.DeepIrisNet2：learning Deep-IrisCodes from scratch for segmentation-robust visible wavelength and near infrared iris recognition[J].arXiv：1902.05390，2019.
[9] NGUYEN K，FOOKES C，ROSS A，et al.Iris recognition with off-the-shelf CNN features：a deep learning perspective[J].IEEE Access，2017，6：18848-18855.
[10] ?PETLíK R，RAZUMENI? I.Iris verification with convolutional neural network and unit-circle layer[C]//41st German Conference on Pattern Recognition.Cham：Springer，2019：274-287.
[11] BOYD A，CZAJKAA，BOWYER K.Deep learning-based feature extraction in iris recognition：use existing models，fine-tune or train from scratch?[J].arXiv：2002.08916，2020.
[12] WANG F，CHENG J，LIU W，et al.Additive margin softmax for face verification[J].IEEE Signal Processing Letters，2018，25（7）：926-930.
[14] HE K，ZHANG X，REN S，et al.Identity mappings in deep residual networks[C]//14th European Conference on Computer Vision.Cham：Springer，2016：630-645.
[15] MISRA D.Mish：a self regularized non-monotonic neural activation function[J].arXiv：1908.08681v1，2019.
[16] PEREYRA G，TUCKER G，CHOROWSKI J，et al.Regularizing neural networksby penalizing confident output distributions[J].arXiv：1701.06548，2017.
[17] 中国科学院自动化研究所.CASIA.v4数据库[DB/OL].[2020-10-20].http：//biometrics.idealtest.org.
Institute of Automation，Chinese Academy of Science.CASIA.v4 database[DB/OL].[2020-10-20].http：//biometrics.idealtest.org.
[18] KUMAR A，PASSI A.Comparison and combination of iris matchers for reliable personal authentication[J].Pattern Recognition，2010，43（3）：1016-1026.
[19] OTHMAN N，DORIZZI B，GARCIA-SALICETTI S.OSIRIS：an open source iris recognition software[J].Pattern Recognition Letters，2016，82：124-131.
[20] SELVARAJU R R，COGSWELL M，DAS A，et al.Grad-CAM：visual explanations from deep networks via gradient-based localization[J].International Journal of Computer Vision，2020，128（2）：336-359.
[21] LIU N，ZHANG M，LI H，et al.DeepIris：learning pairwise filter bank for heterogeneous iris verification[J].Pattern Recognition Letters，2015，82（2）：154-161.
[22] ZHAO Z，KUMAR A.Towards more accurate iris recognition using deeply learned spatially corresponding features[C]//2017 IEEE International Conference on Computer Vision，2017：3829-3838.
[22] NGUYEN K，FOOKES C，SRIDHARAN S.Constrained design of deep iris networks[J].IEEE Transactions on Image Processing，2020，29：7166-7175.
[23] 晁静静，沈文忠，宋天舒.基于HOG和SVM的双眼虹膜图像的人眼定位算法[J].计算机工程与应用，2019，55（9）：190-195.
CHAO J J，SHEN W Z，SONG T S.Human eye location algorithm based on HOG and SVM for binocular iris image[J].Computer Engineering and Applications，2019，55（9）：190-195.

[1]	陈金鑫，沈文忠. 基于EL-YOLO的虹膜图像人眼定位及分类算法[J]. 计算机工程与应用, 2021, 57(17): 217-223.
[2]	周锐烨，沈文忠. PI-Unet：异质虹膜精确分割神经网络模型的研究[J]. 计算机工程与应用, 2021, 57(15): 223-229.
[3]	彭小红，梁子祥，张军，陈荣发. 深度学习驱动的水下图像预处理研究进展综述[J]. 计算机工程与应用, 2021, 57(13): 43-54.
[4]	刘俊涛，王素娟，林晓明，刘祚时，谭俭辉，宋丹. 结合FCM聚类和边缘感知模型的眼底渗出物检测[J]. 计算机工程与应用, 2020, 56(11): 192-199.
[5]	晁静静，沈文忠，宋天舒. 基于HOG和SVM的双眼虹膜图像的人眼定位算法[J]. 计算机工程与应用, 2019, 55(9): 184-189.
[6]	杨霞，朱晓冬，刘元宁，冯家凯，刘帅. 分块小波特征结合BP神经网络的虹膜识别方法[J]. 计算机工程与应用, 2019, 55(18): 132-139.
[7]	邓玉波，苏娟. 基于纹理方向能量特征的虹膜识别算法[J]. 计算机工程与应用, 2017, 53(15): 196-199.
[8]	黄煜，黄文艳，张婉柔，马文华. 基于面阵CCD的迈克尔逊干涉环计数方法[J]. 计算机工程与应用, 2015, 51(9): 180-184.
[9]	庞建华，姚鹏. 基于Adaboost算法的虹膜合格状态检测[J]. 计算机工程与应用, 2015, 51(14): 191-197.
[10]	杨斌，宋立新. 基于自适应区域生长的乳腺肿块分割方法[J]. 计算机工程与应用, 2014, 50(20): 171-175.
[11]	冯莹莹，余世干，刘辉. 改进典型相关分析的虹膜鉴别算法[J]. 计算机工程与应用, 2014, 50(16): 146-149.
[12]	柳培忠1，宁欣2，李卫军2. 一种网页图像文字分割方法[J]. 计算机工程与应用, 2014, 50(10): 180-183.
[13]	贺遵亮，蔡碧野，李峰，章登勇. 采用有向梯度与RANSAC的虹膜定位算法[J]. 计算机工程与应用, 2013, 49(24): 172-174.
[14]	马义德，周炜超，滕飞，绽琨，赵荣昌. 基于SCM的虹膜干扰区域去除算法[J]. 计算机工程与应用, 2013, 49(20): 161-166.
[15]	吴丽媛，徐国华，余琨. 基于前视声纳的成像与多目标特征提取[J]. 计算机工程与应用, 2013, 49(2): 222-225.