中文摘要：

步态识别的准确性容易受到衣着类型及携带背包等局部变化的影响。针对这一问题,首先提出一种基于局部信息熵值的子模式划分方法;然后对正常行走和局部变化两种状态下的每一对子特征进行典型相关分析,得到多个最佳投影矩阵对,并将子特征分别投影到基于上述最佳投影矩阵对的特征子空间中;最后以整体相关系数作为分类依据,以减小局部变化对于整体识别结果的影响。在CASIA-B数据库上的实验表明在所有视角下所提算法都能取得较好的性能。

英文摘要：

Gait recognition would be greatly affected by some covariate factors including clothing type and carrying objects.Finding an approach robust to these covariate factors is the most challenging problem. In this paper, we propose a method based on canonical correlation analysis（CCA） to model the correlation between gait sequences on two different walking conditions. GEIs are partitioned into several parts based on local information entropy value, with each part selected as a sub-pattern.Each pair of sub-pattern are projected onto two learned feature subspaces in which the two transformed gait data sets are optimally correlated based on CCA. Finally, to reduce the effect of the covariate factors, overall correlation strength is used as similarity measure. Experiment results on CASIA-B gait database show that our proposed method outperforms other classical methods over all views.