融合局部特征与深度学习的三维掌纹识别  

3D palmprint recognition by using local features and deep learning

作　　者：杨冰[1] 莫文博 姚金良[1] YANG Bing;MO Wen-bo;YAO Jin-liang(School of Computer Science and Technology,Hangzhou Dianzi University,Hangzhou 310018,China)

机构地区：[1]杭州电子科技大学计算机学院,浙江杭州310018

出　　处：《浙江大学学报（工学版）》2020年第3期540-545,共6页Journal of Zhejiang University：Engineering Science

基　　金：国家自然科学基金资助项目(61402143).

摘　　要：为了探索三维掌纹在生物特征识别领域的应用,基于局部纹理特征和深度学习,提出一种有效的三维掌纹识别方法.通过曲率特征、形状指数、表面类型分别来描述三维掌纹的局部几何特征,将其作为深度神经网络的输入,完成三维掌纹识别任务.在香港理工大学的三维掌纹数据库上对不同的几何特征、不同的深度神经网络模型进行全面分析与比较.三维掌纹识别实验结果表明,与其他三维掌纹识别方法相比较,所提方法的识别率更高,识别时间更短,在实时掌纹识别领域具有较大的应用潜力.An efficient 3 D palmprint recognition method was proposed by using local texture feature sets and deep learning, in order to explore the usage of 3 D palmprint in biometrics recognition. Curvature feature, shape index and surface type were employed to describe the geometry characteristics of local regions in 3 D palmprint data, and then take the charasteristics as the input of the deep neural network to finish 3 D palmprint recognition task.Comprehensive experiments on Hong Kong Polytechnic University 3 D palmprint database were further conducted by using different geometrical features and deep neural network models. The final experimental results of 3 D palmprint recognition validate that the proposed method outperforms existing state-of-the-art methods in terms of recognition accuracy and runtime, showing high potential for real-time palmprint recognition applications.

关 键 词：三维掌纹 局部几何特征 曲率特征 形状指数 表面类型 深度学习 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]