基于侧链连接卷积神经网络的手掌静脉图像识别

doi:10.11772/j.issn.1001-9081.2020050667

计算机应用 ›› 2020, Vol. 40 ›› Issue (12): 3673-3678.DOI: 10.11772/j.issn.1001-9081.2020050667

• 虚拟现实与多媒体计算 • 上一篇下一篇

基于侧链连接卷积神经网络的手掌静脉图像识别

娄梦莹, 王天景, 刘娅琴, 杨丰, 黄靖

南方医科大学生物医学工程学院, 广州 510515

收稿日期:2020-05-19 修回日期:2020-07-21 出版日期:2020-12-10 发布日期:2020-08-14
通讯作者: 刘娅琴(1965-),女,湖南邵阳人,教授,硕士,主要研究方向:生物特征识别、图像处理。liuyq@smu.edu.cn
作者简介:娄梦莹(1994-),女,山东滨州人,硕士研究生,主要研究方向:图像处理、模式识别;王天景(1998-),男,江西九江人,主要研究方向:图像处理、模式识别;杨丰(1965-),男,湖北麻城人,教授,博士,主要研究方向:医学成像技术、机器学习;黄靖(1981-),男,湖南长沙人,副教授,博士,主要研究方向:生物特征识别、图像处理
基金资助:
国家自然科学基金资助项目（61771233）；广东省研发与产业化项目（2013B090500104）。

Palm vein image recognition based on side chain connected convolution neural network

LOU Mengying, WANG Tianjing, LIU Yaqin, YANG Feng, HUANG Jing

School of Biomedical Engineering, Southern Medical University, Guangzhou Guangdong 510515, China

Received:2020-05-19 Revised:2020-07-21 Online:2020-12-10 Published:2020-08-14
Supported by:
This work is partially supported by the National Natural Science Foundation of China （61771233）， the Research and Development and Industrialization Project of Guangdong （2013B090500104）.

摘要/Abstract

摘要： 针对手掌静脉图像数量少且质量参差不齐，进而导致掌脉识别系统的性能降低的现象，提出一种基于侧链连接卷积神经网络的手掌静脉图像识别方法。首先，在ResNet模型的基础上，用卷积层和池化层提取掌脉特征。然后，采用指数线性单元（ELU）激活函数、批归一化（BN）和Dropout技术来改进和优化模型，以缓解梯度消失、防止过拟合、加快收敛及增强模型泛化能力。最后，引入稠密连接网络（DenseNet），使提取到的手掌静脉特征更具丰富性和有效性。在两个公开库和一个自建库上分别进行实验，结果表明所提方法在三个数据库上的识别率分别为99.98%、97.95%、97.96%。可见该方法能有效提高掌脉识别系统的性能，且更适用于掌脉识别的实际应用。

关键词: 手掌静脉识别, ResNet, 指数线性单元激活函数, 批归一化, Dropout, 稠密连接网络

Abstract: To overcome the performance degradation of palm vein recognition system due to the small quantity and the uneven quality of palm vein images, a palm vein image recognition method based on side chain connected convolutional neural network was proposed. Firstly, palm vein features were extracted by convolution layer and pooling layer based on ResNet model. Secondly, the Exponential Linear Unit (ELU) activation function, Batch Normalization (BN) and Dropout technology were used to improve and optimize the model, so as to alleviate gradient disappear, prevent over fitting, speed up convergence and enhance the generalization ability of the model. Finally, Densely Connected Network (DenseNet) was introduced to make the extracted palm vein features more abundant and effective. Experimental results on two public databases and one self-built database show that, the recognition rates of the proposed method on the three databases are 99.98%, 97.95%, 97.96% respectively, indicating that the proposed method can effectively improve the performance of palm vein recognition system, and is more suitable for the practical applications of palm vein recognition.

Key words: palm vein recognition, ResNet, Exponential Linear Unit (ELU) activation function, Batch Normalization (BN), Dropout, Densely Connected Network (DenseNet)

中图分类号:

TP391.4

娄梦莹, 王天景, 刘娅琴, 杨丰, 黄靖. 基于侧链连接卷积神经网络的手掌静脉图像识别[J]. 计算机应用, 2020, 40(12): 3673-3678.

LOU Mengying, WANG Tianjing, LIU Yaqin, YANG Feng, HUANG Jing. Palm vein image recognition based on side chain connected convolution neural network[J]. Journal of Computer Applications, 2020, 40(12): 3673-3678.

参考文献

[1] 吴微, 苑玮琦. 手掌静脉图像识别技术综述[J]. 中国图象图形学报, 2013, 18(10):1215-1224.(WU W,YUAN W Q. A review of palm-vein image recognition[J]. Journal of Image and Graphics, 2013,18(10):1215-1224.)
[2] KAVITHA S,SRIPRIYA P. A review on palm vein biometrics[J]. International Journal of Engineering and Technology, 2018, 7(3.6):407-409.
[3] 徐笑宇, 姚鹏. 基于HOG与改进的SVM的手掌静脉识别算法[J]. 计算机工程与应用, 2016, 52(11):175-180, 214.(XU X Y, YAO P. Palm vein recognition algorithm based on HOG and improved SVM[J]. Computer Engineering and Applications, 2016,52(11):175-180,214.)
[4] PAN M,KANG W. Palm vein recognition based on three local invariant feature extraction algorithms[C]//Proceedings of the 2011 Chinese Conference on Biometric Recognition,LNCS 7098. Berlin:Springer,2011:116-124.
[5] KUBANEK M,SMORAWA D,HOLOTYAK T. Feature extraction of palm vein patterns based on two-dimensional density function[C]//Proceedings of the 2015 International Conference on Artificial Intelligence and Soft Computing,LNCS 9120. Cham:Springer, 2015:101-111.
[6] 周宇佳, 刘娅琴, 杨丰, 等. 基于方向特征的手掌静脉识别[J]. 中国图象图形学报, 2014, 19(2):243-252.(ZHOU Y J,LIU Y Q, YANG F,et al. Palm-vein recognition based on oriented features[J]. Journal of Image and Graphics,2014,19(2):243-252.)
[7] ZHANG B,GAO Y,ZHAO S,et al. Local derivative pattern versus local binary pattern:face recognition with high-order local pattern descriptor[J]. IEEE Transactions on Image Processing, 2010,19(2):533-544.
[8] WANG R,WANG G,CHEN Z,et al. A palm vein identification system based on Gabor wavelet features[J]. Neural Computing and Applications,2014,24(1):161-168.
[9] 刘新亮, 李星野. 基于Radon变换的多尺度虹膜特征提取算法[J]. 光电子·激光, 2008, 19(4):532-536.(LIU X L,LI X Y. A multi-scale iris feature extraction algorithm based on radon transform[J]. Journal of Optoelectronics·Laser,2008,19(4):532-536.)
[10] ELNASIR S,SHAMSUDDIN S M,FAROKHI S. Accurate palm vein recognition based on wavelet scattering and spectral regression kernel discriminant analysis[J]. Journal of Electronic Imaging,2015,24(1):Article No. 013031.
[11] SAVITHA A P,RAMEGOWDA. Acomparative study of palm vein feature extraction and classification[J]. Materials Today:Proceedings,2017,4(11):11882-11887.
[12] ELNASIR S,SHAMSUDDIN S M. Proposed scheme for palm vein recognition based on linear discrimination analysis and nearest neighbour classifier[C]//Proceedings of the 2014 International Symposium on Biometrics and Security Technologies. Piscataway:IEEE,2014:67-72.
[13] GUO X,WANG C,PING Z. A palm-vein recognition algorithm based on LPP and HM-LBP[J]. IOP Conference Series Earth and Environmental Science,2018,153(3):Article No. 032027.
[14] 杜东阳, 路利军, 符瑞阳, 等. 手掌静脉识别:基于端到端卷积神经网络方法[J]. 南方医科大学学报, 2019, 39(2):207-214. (DU D Y,LU L J,FU R Y,et al. Palm vein recognition based on end-to-end convolutional neuralnetwork[J]. Journal of Southern Medical University,2019,39(2):207-214.)
[15] HE K,ZHANG X,REN S,et al. Deep residual learning for image recognition[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE, 2016:1-12.
[16] 黄盛, 李菲菲, 陈虬. 基于改进的深度残差网络的计算断层扫描图像分类算法[J]. 光学学报, 2020, 40(3):50-58.(HUANG S, LI F F,CHEN Q. Computed tomography image classification algorithm based on improved deep residualnetwork[J]. Acta Optica Sinica,2020,40(3):50-58.)
[17] CLEVERT D A,UNTERTHINER T,HOCHREITER S. Fast and accurate deepnetwork learning by Exponential Linear Units (ELUs)[EB/OL].[2019-11-20]. https://arxiv.org/pdf/1511.07289.pdf.
[18] IOFFE S,SZEGEDY C. Batch normalization:accelerating deepnetwork training by reducing internal covariate shift[EB/OL].[2019-11-20]. https://arxiv.org/pdf/1502.03167.pdf.
[19] HINTON G E, SRIVASTAVA N, KRIZHEVSKY A, et al. Improving neuralnetworks by preventing co-adaptation of feature detectors[EB/OL].[2019-11-20]. https://arxiv.org/pdf/1207.0580.pdf.
[20] HUANG G,LIU Z,VAN DER MAATEN L,et al. Densely connected convolutionalnetworks[C]//Proceedings of the 2017 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE,2017:2261-2269.
[21] 娄梦莹, 袁丽莎, 刘娅琴, 等. 基于自适应融合的手掌静脉增强方法[J]. 计算机应用, 2019, 39(4):1176-1182.(LOU M Y, YUAN L S,LIU Y Q,et al. Palm vein enhancement method based on adaptive fusion[J]. Journal of Computer Applications, 2019,39(4):1176-1182.)
[22] LOU M Y,LIU Y Q,YANG F,et al. Image enhancement of palm veins based on adaptive fusion and Gabor filter[C]//Proceedings of the 20195nd International Conference on Fuzzy Systems and Data Mining. Amsterdam:IOS Press,2019:296-304.
[23] QIU S,LIU Y,ZHOU Y,et al. Finger-vein recognition based on dual-sliding window localization and pseudo-elliptical transformer[J]. Expert Systems with Applications,2016,64:618-632.
[24] KRIZHEVSKY A,SUTSKEVER I,HINTON G E. ImageNet classification with deep convolutional neuralnetworks[C]//Proceedings of the 201225th International Conference on Neural Information Processing Systems. Red hook:Curran Associates Inc.,2012:1097-1105.
[25] SZEGEDY C, LIU W, JIA Y, et al. Going deeper with convolutions[C]//Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE. 2015:1-9.

基于侧链连接卷积神经网络的手掌静脉图像识别

Palm vein image recognition based on side chain connected convolution neural network

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 5

编辑推荐

Metrics

[1]	张京爱, 王江涛. 基于多尺度卷积神经网络和类内mixup操作的磁瓦表面质量识别[J]. 计算机应用, 2021, 41(1): 275-279.
[2]	程俊华, 曾国辉, 鲁敦科, 黄勃. 基于Dropout的改进卷积神经网络模型平均方法[J]. 计算机应用, 2019, 39(6): 1601-1606.
[3]	王建仁, 马鑫, 段刚龙, 薛宏全. 边缘智能背景下的手写数字识别[J]. 计算机应用, 2019, 39(12): 3548-3555.
[4]	刘雨桐, 李志清, 杨晓玲. 改进卷积神经网络在遥感图像分类中的应用[J]. 计算机应用, 2018, 38(4): 949-954.
[5]	沈承恩, 何军, 邓扬. 基于改进堆叠自动编码机的垃圾邮件分类[J]. 计算机应用, 2016, 36(1): 158-162.