Palm Vein Recognition Network Combining Transformer and CNN

doi:10.3778/j.issn.1002-8331.2208-0086

Abstract

Abstract: Aiming at the low accuracy of palm vein feature extraction and recognition, it proposes a palm vein recognition network PVCodeNet. It designs an improved BasicBlock and Transformer Encoder, and uses AAM-loss（additional angular margin loss） to expand decision boundary. It successfully applies Transformer Encoder to global feature extraction of palm vein firstly. Improved BasicBlock uses Do-Conv to replace Conv for feature extraction, it makes extracted features more distinctive. it also adds standardized attention module NAM, its detailed features of in channel and spatial domain are extracted by applying heavy sparsity penalty to suppress weights of insignificant features. This paper describes in detail the key point location, ROI extraction and image enhancement, then makes detailed experiments on feature vector dimension and AAM-loss parameter setting. Finally, ablation experiments are carried out on PolyU database and selfbuilt database SEPAD-PV, EER reaches 0, it achieves a breakthrough in the highest recognition rate. In order to verify the generalization performance of network, it is also verifies on the palmprint database Tongji and the finger vein database SDUMLA with similar texture features. EER is far superior to other mainstream algorithms, which fully proves the superiority of this algorithm.

Key words: palm vein recognition, Transformer encoder, Do-Conv, normalization-based attention module（NAM）, additive angular margin loss（AAM-Loss）

摘要： 针对手掌静脉特征提取识别精度不高问题，提出了掌静脉识别网络PVCodeNet。该网络设计了改进的BasicBlock和Transformer Encoder模块结合并运用扩大决策边界的损失函数AAM-Loss（additive angular margin loss）。该网络首次将Transformer Encoder模块成功用于掌静脉图像全局特征提取，改进的BasicBlock使用深度超参数化卷积Do-Conv取代传统卷积Conv进行特征提取使提取的特征更加具有区分性，该模块还加入规一化的注意力机制NAM模块，通过应用权重稀疏性惩罚项抑制不显著性特征的权值来提取图像在通道和空间域上重要的细节特征。在手掌关键点定位、ROI提取、图像增强方面作了详细描述，在特征向量维度、AAM-Loss参数设置方面做了详细实验，在PolyU数据库和自建库SEPAD-PV数据库上进行消融实验测试，EER均达到了0，成功实现了最高识别率的突破。为了验证该网络的泛化性能，还在具有相似纹理特征的掌纹数据库Tongji和指静脉数据库SDUMLA上进行验证，EER远远优于其他主流算法，充分证明了提出算法的优越性。

关键词: 手掌静脉识别, Transformer编码模块, 深度超参数化卷积（Do-Conv）, 规一化注意力机制（NAM）, 扩大决策边界的损失函数（AAM-Loss）

WU Kai, SHEN Wenzhong, JIA Dingding, LIANG Juan. Palm Vein Recognition Network Combining Transformer and CNN[J]. Computer Engineering and Applications, 2023, 59(24): 98-109.

吴凯, 沈文忠, 贾丁丁, 梁娟. 融合Transformer和CNN的手掌静脉识别网络[J]. 计算机工程与应用, 2023, 59(24): 98-109.

References

[1] LU G?M，LI?H?B，LIU?L.A review of biometric recognition[J].Nature Science，2007，27（1）：81-87.
[2] WANG?Y?H，TAN?T?N.Modern identity identification new technology biometric technology[J].Chinese Basic Science，2010，1（9）：6-12.
[3] SONG?D，HUANG X.Biometric technology in field of financial payment security[J].Information Security Research，2016，2（1）：27-32.
[4] LADOUX P，ROSENBERGER C，DORIZZI B，et al.Palm vein verification system based on SIFT matching[C]//International Conference on Biometrics.Biometr：Springer，2009：1290-1298.
[5] LEE Y P.Palm vein recognition based on a modified（2D） LDA[J].Signal Image Video Process，2015，9（1）：229-242.
[6] WANG J，YANG W Y，SU A.Person recognition by fusing palmprint and palm vein images based on “Laplacianpalm” presentation[J].Pattern Recognition，2008，41（5）：1514-1527.
[7] LU W J，LI M N，ZHANG L.Palm vein recognition using directional features derived from local binary patterns[J].International Journal of Signal Processing，Image Processing and Pattern Recognition，2016，9（5）：87-98.
[8] ZHOU Y J，LIU Y Q，FENG Q.Palm vein classification based on principal location features[J].PLoS One，2014，9（11）：e112429.
[9] ZHANG D，GUO Z H，LU G M.An online system of multi spectral palmprint verification[J].IEEE Transactions on Instrumentation and Measurement，2010，59（2）：480-490.
[10] SHI Y H，YANG J F.Image restoration and enhancement for the finger vein recognition[C]//2012 IEEE 1st International Conference on Signal Processing，October 21-25，2012，Beijing，China.New York：IEEE Press，2012：1605.
[11] WU J D，LIU C T.Finger-vein pattern identification using SVM and neural network technique[J].Expert Systems with Applications，2010，38（11）：14284-14289.
[12] RAGHAVENDRA R，SURBIRYALA J，BUSCH C.An efficient finger vein indexing scheme based on unsupervised clustering[C]//IEEE International Conference on Identity，Security and Behavior Analysis（ISBA 2015），March 23-25，2015，Hong Kong，China.New York：IEEE Press，2015：1-8.
[13] ZHAO X H，YIN L F，ZHU Y N.Improved image classification algorithm based on principal component analysis network[J].Laser and Optoelectronics Progress，2016，56（2）：021004.
[14] LI X C，MA H Y，LIN S.Palmprint recognition based on subspace and texture fusion[J].Laser and Optoelectronics Progress，2019，56（7）：071007.
[15] WANG G，SUN C.Multi-weighted co-occurrence descriptor encode for vein recognition[J].IEEE Transactions on Information Forensics and Security，2020，10（8）：375-390.
[16] AHMAD F，CHENG L M.Lightweight and privacy preserving template generation for palm-vein based human recognition[J].IEEE Transactions on Information Forensics and Security，2019，5（12）：1012-1024.
[17] THAPAR D，JASWAL G，NIGAM A.Palm vein authentication siamese network trained using triplet loss and adaptive hard mining by learning enforced domain special features[C]//IEEE 5th International Conference on Identity，Security，and Behavior Analysis（ISBA），2019：865-876.
[18] HE K，ZHANG X，REN S，et al.Deep residual learning for image recognition[C]//IEEE Conference on Computer Vision & Pattern Recognition，2016.
[19] GUO Z H，ZHANG D，ZHANG L.Empirical study of light source selection for palm vein recognition letters[J].IEEE Transactions on Information Forensics and Security，2011，32（2）：120-126.
[20] YANG B，XIANG X Q，XU D Q.Palmprint recognition using shape index representation and fragile bits[J].Multimedia Tools and Applications，2017，76（14）：15357-15375.
[21] YIN Y L，LIU L，SUN X.SDUMLA-HMT：a multimodal biometric database[C]//Chinese Conference on Biometric Recognition Springer，2011：260-268.
[22] WANG F，LIU W，LIU H.Additive margin softmax for face verification[J].arXiv：1801.05599，2018.
[23] ALEXEY D，LUCAS B，ALEXANDER K，et al.An image is worth 16x16 words：transformers for image recognition at scale[C]//International Conference on Learning Representations，2020：9-29.
[24] LIU Z，LIN Y T，CAO Y.Swin transformer：hierarchical vision transformer using shifted windows[C]//International Conference on Learning Representations，2021：2103-2130.
[25] LIU Z，MAO H Z，WU C Y，et al.A ConvNext for the 2020s[C]//International Conference on Learning Representations，2022：527-545.
[26] ZHOU Y B，KUMAR A.Human identification using palm vein images[J].IEEE Transactions on Information Forensics and Security，2019，11（4）：1259-1274.
[27] ZHANG D，GUO Z H，LU G M.An online system of multispectral palmprint verification[J].IEEE Transactions on Instrumentation and Measurement，2019，2（9）：480-490.
[28] KUZU R S，MAIORANA E，CAMPISI P.On-the-fly finger-vein-based biometric recognition using deep neural networks[J].IEEE Transactions on Information Forensics and Security，2020，24（15）：2641-2654.
[29] HORNG S J，VU D T，NGUYEN T V，et al.Recognizing palm vein in smartphones using RGB images[J].IEEE Transactions on Informatics，2022，18（9）：1904-1916.
[30] SHAO H，ZHONG D，DU X，et al.Few-shot learning for palmprint recognition via meta-siamese network[J].IEEE Transactions on Instrumentation and Measurement，2021，70（1）：1-12.
[31] EDWIN S，RUBER H G，RICARDO J，et al.Generating style-based palm vein synthetic images for the creation of large-scale datasets[C]//11th International Conference Pattern Recognition Systems，2021：182-187.
[32] DAVID Z B，RUBER H G，RICARDO J B，et al.Evaluation of the standard and regularized elms for gender and age classification based on palm vein images[C]//International Conference of the Chilean Computer Science Society，2021：1204-1218.
[33] JIA W，GAO J，XIA W，et al.A performance evaluation of classic convolutional neural networks for 2D and 3D palmprint and palm vein recognition[J].International Journal of Automation and Computing，2021，18（6）：612-628.
[34] BABALOLA F O，BITIRIM Y，TOYGARL O.Palm vein recognition through fusion of texture-based and CNN-based methods[J].Signal，Image and Video Processing，2021，15（7）：459-466.
[35] SUN S，CONG X，ZHANG P，et al.Palm vein recognition based on NPE and KELM[J].IEEE Access，2022，9（15）：71778-71783.