摘要
步态识别是图像处理领域的一个新兴热点。人行走姿态准确识别困难因素较多,由于步态数据是一种高维、小样本数据,传统识别方法不能检测前景与背景差异情况,导致识别正确率比较低。为了快速准确地进行步态识别,提出支持向量机的步态识别方法。方法首先根据步态图像中前景点与背景点的差值,自适应计算区分前景点与背景点的阈值,根据阈值对步态图像进行二值化,在特征提取阶段,采用水平、垂直和对角线3个方向提取步态信息,并通过小波变换进行特征维数约简,最后将小波变换提取维步态特征采用支持向量机学习得到步态识别结果。在中国科学院自动化所的CASIA步态数据库上进行了识别仿真,结果表明,方法的识别正确率有所提高,且识别的速度加快,是步态识别有效的方法,并具有广阔的应用前景。
Gait recognition is a new field of image processing.Classic gait identification method can not detect foreground and background differences and is difficult to obtain good generalization performance.In order to impove the recognition rate,a gait recognition method is proposed based on support vector machine(SVM).This method uses adaptive calculation to distinguish the foreground and background points threshold firstly,and then adopts horizontal and vertical direction and diagonal,and through information extraction of wavelet transform characteristic dimension reduction,finally the gait characteristic extracted with wavelet transform is learned by using support vector machine to obtain the gait recognition results.With the CASIA gait database of Chinese academy of sciences institute of automation,recognition simulation experiments are carried out,The results show that this method improves the accuracy and speed of recognition,is one of the effective method in gait identification field,and has wide application prospects.
出处
《计算机仿真》
CSCD
北大核心
2011年第3期302-305,398,共5页
Computer Simulation
关键词
步态识别
小波变换
支持向量机
Gait recognition
Wavelet transform
Support vector machine(SVM)