特征融合的多视角步态识别研究  

Research on multi-perspective gait recognition using feature fusion

作　　者：王竣[1] 王修晖[1] 

机构地区：[1]中国计量大学信息工程学院,浙江杭州310018

出　　处：《中国计量大学学报》2017年第2期234-240,268,共8页Journal of China University of Metrology

基　　金：国家自然科学基金资助项目(No.61303146)

摘　　要：针对单一特征步态识别率低的问题,提出一种将步态能量图(Gait Energy Image,GEI)中动态部分和Gabor小波特征融合的步态识别算法.首先,通过运动目标检测及二值化和形态学处理等预处理操作得到步态轮廓图,再进一步从步态轮廓图计算得到步态能量图,并从中分割出动态部分.然后,利用Gabor小波从步态能量图的动态部分中提取不同角度的信息,将两步态特征融合在一起,对融合后得到的特征向量用改进的KPCA方法进行降维.最后,将降维后的融合特征向量输入到基于多分类的支持向量机(Support Vector Machine,SVM)中,从而完成步态的分类和识别.经过在中国科学院自动化研究所CASIA步态数据库上进行实验,取得了很好的识别效果,实验结果表明,与单一特征的步态识别方法相比,融合后算法的识别率提高了约10%.Since the low gait recognition rate of single features, a novel gait recognition algorithm based on the feature fusion of the dynamic part of gait energy images （GEI） and Gabor was presented in this paper. Firstly, the gait contour images were extracted through the object detection, binarization and morphological process, to caluclate GEI and divide the dynamic part of GEL Secondly, the information of different angles was extracted from the dynamic part of GEI with Gabor wavelets. Feature fusion was extracted with GEI and Gabor wavelets and was reduced by improved KPCA. Finally, the vectors of feature fusion were input into the SVM （Support Vector Machine） based on multi classification to realize the classification and recognition of gait. Experiments were conducted on the Central Asia Student International Academic（CASIA） gait database with satisfactory recognition effect. Compared with methods based on the single gait feature, the gait recognition rate after the fusion was executed increased about 10%.

关 键 词：GABOR小波 步态能量图 特征融合 改进的KPCA 支持向量机 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]