| 691 | 7 | 11 |
| 下载次数 | 被引频次 | 阅读次数 |
针对人脸识别技术应用中，使用人脸图像或虚拟人脸等技术手段欺骗识别系统进而降低系统安全性的问题，提出了一种多模态特征融合的人脸活体检测算法。该算法将人脸的RGB图、深度图和红外图分别输入到3个相同的残差网络结构中提取特征；通过基于通道注意力机制的方式对3个模态的特征图进行融合；对融合特征做出决策。在CASIA-SURF数据集上多次实验表明，人脸活体检测的准确率为99.9%,能高效区分出真实人脸和伪造人脸。
Abstract：To solve the problem of using face image or virtual face deception recognition system to reduce system security when applying face recognition technology, a multi-mode feature fusion face anti-spoofing algorithm is proposed.In this algorithm, the RGB image, depth image and infrared image of human face are input respectively into three identical residual network structures for feature extraction.The feature maps of the three modes are fused based on the channel attention mechanism, and a decision is made on the fusion feature.Multiple experiments on the CASIA-SURF dataset show that the accuracy of face anti-spoofing is 99.9%,which is high enough to distinguish real faces from fake faces effectively.
[1] ZHANG L,PENG F,QIN L,et al.Face Spoofing Detection Based on Color Texture Markov Feature and Support Vector Machine Recursive Feature Elimination[J].Journal of Visual Communication and Image Representation,2018,51:56-69.
[2] SINGH A K,JOSHI P,NANDI G C.Face Recognition with Liveness Detection Using Eye and Mouth Movement[C]//International Conference on Signal Propagation and Computer Technology.Ajmer:IEEE,2014:592-597.
[3] BHATTACHARJEE S,MOHAMMADI A,MARCEL S.Spoofing Deep Face Recognition with Custom Silicone Masks[C]//9th IEEE International Conference on Biometrics Theory,Applications and Systems.Long Beach:IEEE,2018:1-7.
[4] LIU Y,JOURABLOO A,LIU X.Learning Deep Models for Face Anti-spoofing:Binary or Auxiliary Supervision[C]//IEEE/CVF Conference on Computer Vision and Pattern Recognition.Washington DC:IEEE,2018:389-398.
[5] 邓雄，王洪春.基于深度学习和特征融合的人脸活体检测算法[J].计算机应用，2020,40(4):1009-1015.
[6] 赵洋，许军.基于MobileNetV2与树莓派的人脸识别系统[J].计算机系统应用，2021,30(8):67-72.
[7] 聂建军，刘书洪，叶少斌.基于多模型融合的可见光人脸活体检测方法[C]//2020年“网络安全技术与应用创新”研讨会.上海：公安部第三研究所《信息网络安全》北京编辑部，2020:1-4.
[8] 吴晓丽，胡伟.基于注意力的热点块和显著像素卷积神经网络的人脸防伪方法[J].计算机科学，2021,48(4):316-324.
[9] HE K,ZHANG X,REN S,et al.Deep Residual Learning for Image Recognition[C]//IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Las Vegas:IEEE,2016:770-778.
[10] 李立鹏，师菲蓬，田文博，等.基于残差网络和迁移学习的野生植物图像识别方法[J].无线电工程，2021,51(9):857-863.
[11] ZHANG X,ZHAO R,QIAO Y,et al.AdaCos:Adaptively Scaling Cosine Logits for Effectively Learning Deep Face Representations[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Long Beach:IEEE,2019:10815-10824.
[12] 殷亚萍，柴文，凌毅德，等.基于特征分析的卷积神经网络烟雾识别[J].无线电工程，2021,51(7):526-533.
[13] 卫依雪，周冬明，王长城，等.采用多尺度密集残差网络的水下图像增强[J].无线电工程，2021,51(9):870-878.
[14] HU J,SHEN L,ALBANIE S,et al.Squeeze-and-excitation Networks[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2020,42(8):2011-2023.
[15] 江培营，陶青川，艾梦琴.基于注意力机制和深度学习的钢板表面缺陷图像分类[J].计算机应用与软件，2021,38(9):214-219.
[16] LOSHCHILOV I,HUTTER F.SGDR:Stochastic Gradient Descent with Warm Restarts[C]//5th International Conference on Learning Representations (ICLR 2017).Toulon:IEEE,2016:1-16.
[17] 皮家甜，杨杰之，杨琳希，等.基于多模态特征融合的轻量级人脸活体检测方法[J].计算机应用，2020,40(12):3658-3665.
[18] 孙文赟，金忠，赵海涛，等.基于深度特征增广的跨域小样本人脸欺诈检测算法[J].计算机科学，2021,48(2):330-336.
[19] ZHANG S,WANG X,LIU A,et al.A Dataset and Benchmark for Large-scale Multi-modal Face Anti-spoofing[C]∥2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Long Beach:IEEE,2019:919-928.
[20] SIMONYAN K,ZISSERMAN A.Very Deep Convolution Networks for Large-scale Image Recognition[J/OL].[2015-04-10].https://arxiv.org/abs/1409.1556.
[21] MA N,ZHANG X,ZHENG H,et al.ShuffleNet V2:Practical Guidelines for Efficient CNN Architecture Design[C]//2018 European Conference on Computer Vision.Munich:Springer,2018:122-138.
[22] SANDLER M,HOWARD A,ZHU M,et al.MobileNetV2:Inverted Residuals and Linear Bottlenecks[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.Salt Lake City:IEEE,2018:4510-4520.
[23] 隋孟君，茅耀斌，孙金生.基于近红外图像特征的活体人脸检测[J].自动化与仪器仪表，2021,9:25-29.
[24] MOHAMED S,GHONEIM A S,YOUSSIF A.Visible/Infrared Face Spoofing Detection Using Texture Descriptors[J].MATEC Web of Conferences,2019,292(2):1-5.
[25] 黄俊，张娜娜，章惠.融合头部姿态和面部表情的互动式活体检测[J].计算机应用，2020,40(7):2089-2095.
基本信息:
DOI：
中图分类号:TP391.41
引用信息:
[1]赵洋,许军,靳永强等.多模态特征融合的人脸活体检测算法[J].无线电工程,2022,52(05):738-744.
基金信息:
辽宁省自然基金项目(2019-ZD-0068); 辽宁省教育厅项目(XXLJ2019010)~~