基于SEF-GHEI及协同表示的步态识别研究

将协同表示方法应用于步态识别中可以解决稀疏表示方法计算耗时的问题,但提取步态特征采用的GEI算法没有考虑步态内部轮廓边界信息,导致识别率不高。针对此问题,本文提出使用融合Hog和GEI算法的方法提取步态特征,在此基础上使用协同表示的方法训练,再通过计算测试样本的最小重构误差进行分类。实验结果表明,该方法在单一视角下步态识别准确率平均提高了1.315%,以及跨视角下步态识别准确率平均提高了6.51%,说明本方法是可行的。

Research on a gait recognition based on SEF-GHEI and collaborative representation

Applying collaborative representation method to gait recognition research can solve the problem of time-consuming calculation of sparse representation. While the GEI algorithm for extracting gait features does not consider gait internal outline information, resulting a low recognition rate. To address this problem, this article propose a novel method to extract gait features by using a fusion Hog and GEI algorithm. On the basis of this method, train the samples by using collaborative representation method and classify the samples according to the minimum reconstruction error. The experimental results show that the accuracy of the proposed method is 1.315% higher than that of the collaborative representation in a single view, and the accuracy of gait recognition is increased by 6.51% over the cross view.