摘要
目的:解决人工步态特征参数模型识别率低的问题。方法:采用基于CNN网络结构和SVM分类器的步态识别方法,提出三种不同CNN网络结构,其中两种在传统CNN网络结构中添加了多通道卷积技术。结果:通过在中国科学院提供的CASIA和日本大阪大学提供的OU-ISIR步态数据库上进行测试和验证,结果显示采用三种CNN网络结构得到的步态识别率都有显著提高,其中晚期多通道卷积CNN网络结构(LCNN)得到的步态识别率最高。结论:基于卷积神经网络(CNN)的步态识别方法较好地解决了人工步态特征模型造成的低识别率问题。
Aims:This paper aims to solve the problem of low recognition rate of artificial gait feature parameter models.Methods:Three convolutional neural networks were proposed by using a novel gait recognition algorithm based on CNN and SVM,two of which were adopted the multi-channel convolution technology on the traditional CNN network structure.Results:Three experiments were conducted on the CASIA gait database provided by the Institute of Automation of the Chinese Academy of Sciences and the OU-ISIR gait database provided by Osaka University separately.The results showed that the late multi-channel convolution CNN network structure (LCNN) had the highest gait recognition rate.Conclusions:The proposed gait recognition method based on convolutional neural network (CNN) can solve the problem of low recognition rate of artificial gait feature parameter models.
作者
张加加
王修晖
ZHANG Jiajia;WANG Xiuhui(College of Information Engineering,China Jiliang University,Hangzhou 310018,China)
出处
《中国计量大学学报》
2019年第1期65-71,共7页
Journal of China University of Metrology
基金
国家自然科学基金项目(No.61303146
61602431)
