基于深度学习的虹膜人脸多特征融合识别  

Iris-face multi-biometric fusion recognition using deep learning

作　　者：肖珂[1] 汪训昌 何云华[1] 李超飞 田玉通 XIAO Ke;WANG Xun-chang;HE Yun-hua;LI Chao-fei;TIAN Yu-tong(School of Information Science and Technology,North China University of Technology,Beijing 100144,China)

机构地区：[1]北方工业大学信息学院,北京100144

出　　处：《计算机工程与设计》2020年第4期1070-1073,共4页Computer Engineering and Design

基　　金：国家自然科学基金青年基金项目(61802005);国家重点研发基金项目(2017YFB0802300)。

摘　　要：传统多生物特征融合识别方法中人工设计特征提取存在盲目性和差异性,特征融合存在空间不匹配或维度过高等问题,为此提出一种基于深度学习的多生物特征融合识别方法。通过卷积神经网络(convolutional neural networks,CNN)提取人脸和虹膜特征、参数化t-SNE算法特征降维和支持向量机(support vector machine,SVM)分类组合进行融合识别。实验结果表明,该融合识别方法与单一生物特征识别以及其它融合识别方法相比,鲁棒性增强,识别性能提升明显。In the traditional multi-biometric fusion recognition method,there are blindness and difference in artificial design feature extraction and there are some problems in feature fusion such as space mismatch or high dimension.Therefore,a multi-biometric fusion recognition method based on deep learning was proposed.Face and iris features were extracted using convolutional neural networks(CNN),and these extracted features were fused in series and the parametric t-SNE algorithm was used for feature dimensionality reduction.Support vector machine(SVM)was used for classification and recognition.The experimental data show that the fusion recognition method has stronger robustness and better recognition performance than single biometric recognition and other fusion recognition methods.

关 键 词：人脸识别 虹膜识别 多特征融合 卷积神经网络 特征降维 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]