基于改进的VGGNet算法的人脸识别  

Face recognition based on improved VGGNet algorithm

作　　者：喻丽春 刘金清 YU Lichun;LIU Jinqing(Department of Information Technology,Fuzhou University of International Studies and Trade,Fuzhou 350202,China;College of Photonic and Electronic Engineering,Fujian Normal University,Fuzhou 350007,China)

机构地区：[1]福州外语外贸学院信息系,福建福州350202 [2]福建师范大学光电与信息工程学院,福建福州350007

出　　处：《长春工业大学学报》2018年第4期378-385,共8页Journal of Changchun University of Technology

基　　金：福建省自然科学基金资助项目(2010J01327)

摘　　要：基于小规模数据集,在有限计算能力应用场景下进行VGGNet改进人脸识别。实验结果表明,改进后的卷积神经网络对训练样本要求较小,训练速度较快,人脸识别率达95%以上。Based on small scale data set, an improved face recognition method based on VGGNet is proposed under limited computing power application scene. Experimental results show that the improved convolution neural network require less training samples but with faster training speed and more than 95% face recognition rate

关 键 词：深度学习 人脸识别 卷积神经网络 特征提取 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]