Silent Live Face Detection in Near-Infrared Images Based on Optimized LeNet-5
-
摘要: 针对当前交互式活体检测过程繁琐、用户体验性差的问题,提出了一种优化LeNet-5和近红外图像的静默活体检测方法。首先,采用近红外光摄像头构建了一个非活体数据集;其次,通过增大卷积核、增加卷积核数目、引入全局平均池化等方法对LeNet-5进行了优化,构建了一个深层卷积神经网络;最后,将近红外人脸图片输入到模型中实现活体静默活体检测。实验结果表明,所设计的模型在活体检测数据集上有较高的识别率,为99.95%,整个静默活体检测系统的运行速度约为18~22帧/s,在实际应用中鲁棒性较高。Abstract: An improved method of silent liveness detection for LeNet-5 and near-infrared images is proposed to overcome the problem of the interactive liveness detection process and poor user experience. First, a face attack dataset was constructed using a near-infrared camera. Second, the LeNet-5 was optimized by increasing the number of convolution kernels and introducing global average pooling to construct a deep convolutional neural network. Finally, the near-infrared face image is input to the model to realize silent liveness detection. The experimental results show that the proposed model has a higher recognition rate for the liveness detection dataset, reaching 99.95%. The running speed of the silent liveness detection system is approximately 18-22 frames per second, which shows high robustness in practical applications.
-
表 1 模型结构参数
Table 1. Model structure parameters
Layer Name Layer Type Output Size/Strides Kernel Size Input Input layer 128×128×3/- - C1 Convolution 128×128×32/1 7 P1 Max Pooling 64×64×32/2 2 C2 Convolution 64×64×64/1 7 P2 Max Pooling 32×32×64/2 2 C3 Convolution 32×32×128/1 5 P3 Max Pooling 16×16×128/2 2 C4 Convolution 16×16×256/1 5 P4 Max Pooling 8×8×256/2 2 C5 Convolution 8×8×512/1 5 P5 Max Pooling 4×4×512/2 2 GAP GAP 1×1×512/1 4 Softmax Softmax 2/- - 表 2 10折交叉验证结果
Table 2. 10-fold cross-validation results
Category Test Dadaset 1 2 3 4 5 6 7 8 9 10 Liveness 99.97 99.97 99.97 100 100 99.97 99.97 99.97 100 99.95 Non-liveness 99.98 99.95 99.91 99.98 99.88 99.93 99.86 99.93 99.91 99.83 Overall 99.96 99.95 99.94 99.95 99.98 99.95 99.91 99.95 99.96 99.90 表 3 三种算法结果比较
Table 3. Comparison of the results of the three algorithms
Algorithm Accuracy/% Average prediction time for a single picture/ms GPU CPU SVM 96.67 - 4.43 LeNet-5 98.23 2.03 7.57 LeNet_Liveness 99.95 10.77 31.08 表 4 不同文献结果比较
Table 4. Comparison of results from different literature
Detection type Literature Algorithm Equipment Accuracy/% Interactive [2] Head posture + mouth opening and closing detection Visible light camera 99.25 [4] Random emoji commands Visible light camera 95.85 [5] Blink detection + smile detection + open mouth detection VTM camera 97.67 Silent [13] LBP+Gabor+SVM Visible light camera 98.00 [14] SVM+3D point cloud reconstruction+Face key point Binocular camera(Near infrared light+visible light) 99.00 [15] CNN(double-mean pooling +multiple types of activation function) Visible light camera 99.67 This article CNN (LeNet-5 improvements) near-infrared camera(Near infrared light) 99.95 -
[1] Singh A K, Joshi P, Nandi G C. Face recognition with liveness detection using eye and mouth movement[C]//Proceedings of the 2014 International Conference on Signal Propagation and Computer Technology (ICSPCT), IEEE, 2014: 592-597. [2] 张进, 张娜娜. 优化特征提取的互动式人脸活体检测研究[J]. 计算机工程与应用, 2019, 55(13): 193-200. doi: 10.3778/j.issn.1002-8331.1804-0227ZHANG Jin, ZHANG Nana. Research on Interactive Face Detection Based on Optimized Feature Extraction[J]. Computer Engineering and Applications, 2019, 55(13): 193-200. doi: 10.3778/j.issn.1002-8331.1804-0227 [3] V David, A Sanchez. Advanced support vector machines and kernel methods[J]. Neurocomputing, 2003, 55(1/2): 5-20. http://www.onacademic.com/detail/journal_1000035125251010_2624.html [4] Ng E S, Chia Y S. Face verification using temporal affective cues[C]// Proceedings of the 21st International Conference on Pattern Recognition, Piscataway, 2012: 1249-1252. [5] 马钰锡, 谭励, 董旭, 等. 面向VTM的交互式活体检测算法[J]. 计算机工程, 2019, 45(3): 256-261. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC201903043.htmMAYuxi, TAN Li, DONG Xu, et al. Interactive Liveness Detection Algorithm for VTM[J]. Computer Engineering, 2019, 45(3): 256-261. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC201903043.htm [6] Lecun Y, Bottou L. Gradient-based learning applied to document recognition[C]//Proceedings of the IEEE, 1998, 86(11): 2278-2324. [7] 李文宽, 刘培玉, 朱振方, 等. 基于卷积神经网络和贝叶斯分类器的句子分类模型[J]. 计算机应用研究, 2020, 37(2): 333-336, 341. https://www.cnki.com.cn/Article/CJFDTOTAL-JSYJ202002003.htmLI Wenkuan, LIU Peiyu, ZHU Zhenfang, et al. Sentence classification model based on convolution neural network and Bayesian classifier[J]. Application Research of Computers, 2020, 37(2): 333-336, 341. https://www.cnki.com.cn/Article/CJFDTOTAL-JSYJ202002003.htm [8] 程淑红, 周斌. 基于改进CNN的铝轮毂背腔字符识别[J]. 计算机工程, 2019, 45(5): 182-186. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC201905029.htmCHENG Shuhong, ZHOU Bing. Recognition of Characters in Aluminum Wheel Back Cavity Based on Improved Convolution Neural Network[J]. Computer Engineering, 2019, 45(5): 182-186. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJC201905029.htm [9] LIN M, CHEN Q, YAN S. Network In Network[EB/OL]. [2014-03-04]. https://arxiv.org/pdf/1312.4400.pdf. [10] ZHANG B, ZHANG L, ZHANG D, et al. Directional binary code with application to PolyU near-infrared face database[J]. Pattern Recognition Letters, 2010, 31(14): 2337-2344. doi: 10.1016/j.patrec.2010.07.006 [11] ZHANG K, ZHANG Z, LI Z, et al. Joint Face detection and alignment using multitask cascaded Convolutional Networks[J]. IEEE Signal Processing Letters, 2016, 23(10): 1499-1503. doi: 10.1109/LSP.2016.2603342 [12] KINGMA D P, BA J. Adam: a method for stochastic optimization[EB/OL]. [2014-12-22]. https://arxiv.org/pdf/1412.6980v8.pdf. [13] Määttä J, Hadid A, Pietikäinen M. Face spoofing detection from single images using micro-texture analysis[C]//Proceedings of the International Joint Conference on Biometrics, IEEE, 2011: 1-7. [14] 邓茜文, 冯子亮, 邱晨鹏. 基于近红外与可见光双目视觉的活体人脸检测方法[J]. 计算机应用, 2020, 40(7): 2096-2103. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJY202007038.htmDENG Qianwen, FENG Ziliang, QIU Pengchen. Face liveness detection method based on near-infrared and visible binocular vision[J]. Journal of Computer Applications, 2020, 40(7): 2096-2103. https://www.cnki.com.cn/Article/CJFDTOTAL-JSJY202007038.htm [15] 龙敏, 佟越洋. 应用卷积神经网络的人脸活体检测算法研究[J]. 计算机科学与探索, 2018, 12(10): 1658-1670. doi: 10.3778/j.issn.1673-9418.1801009LONG Min, TONG Yueyang. Research on Face Liveness Detection Algorithm Using Convolutional Neural Network[J]. Journal of Frontiers of Computer Science and Technology, 2018, 12(10): 1658-1670. doi: 10.3778/j.issn.1673-9418.1801009