基于深度置信网络的假手指静脉图像检测算法  

Fake Finger Vein Image Detection Based on Deep Belief Network

作　　者：刘霞[1] 秦华锋[2] LIU Xia;QIN Hua-feng(National Research Base of Intelligent Manufacturing;School of Computer Science and Information Engineering,Chongqing Technology and Business University,Chongqing 400067,China)

机构地区：[1]重庆工商大学智能制造服务国际科技合作基地,重庆400067 [2]重庆工商大学计算机科学与信息工程学院,重庆400067

出　　处：《重庆工商大学学报（自然科学版）》2018年第5期32-38,共7页Journal of Chongqing Technology and Business University:Natural Science Edition

基　　金：国家自然科学基金(61402063);重庆市自然科学基金(CSTC2017JCYJAX0002);重庆市自然科学基金(CSTC2018JCYJAX0095)

摘　　要：针对伪造的手指静脉图像能够成功攻击手指静脉识别系统,从而使得其识别系统安全性能大大降低的问题,提出了一种基于深度置信网络的手指静脉防伪检测的方法;通过逐层无监督的学习方法预训练深度网络的权值参数,以及有监督的BP神经网络微调深度网络的权值参数,从而提取到手指静脉图像的特征,用于静脉图像的检测;实验结果证明所提出的手指静脉防伪检测方法能够有效地识别出假手指静脉图像;通过对比性实验研究,发现此方法提高了手指静脉识别系统的安全性能。According to that fake finger vein images can successfully attack finger vein recognition system, making the safety performance significantly reduced, this paper proposes a novel method for finger vein spoofing attack detection based on Deep Belief Network. Through the unsupervised training layer-by-layer to get the weight parameters of the depth network and the supervised BP neural network to fine-tune the weight parameters, we extract the feature of the finger vein image and detect the vein image. Obtained experiment results show that the proposed scheme can effectively identify the fake finger vein image. Through comparative experimental research, we find that the proposed scheme can improve the safety performance of the recognition system.

关 键 词：手指静脉 防伪检测 深度学习 深度置信网络 

分 类 号：TP393.0[自动化与计算机技术—计算机应用技术]