基于加强步态能量图在非规范视角的步态识别研究  

Research of E-GEI based gait identification in informal viewing angle

作　　者：肖文[1] 钟秀锋[1] 

机构地区：[1]汕头大学电子工程系,广东汕头515063

出　　处：《现代电子技术》2012年第6期138-140,共3页Modern Electronics Technique

摘　　要：提出了一种基于加强步态能量图的非规范视角步态识别方法,解决了非规范视角下步态识别难题。视角转换方法将视角统一,采用背景减除法提取人体轮廓,引入步态周期检测方法确定步态周期,根据人体骨架参数模型得到加强步态能量图(E-GEI),最后运用2DPCA方法提取特征向量,并采用最近邻域法分类。实验结果表明:E-GEI在各个视角下比普通的GEI在识别效果要更好。A method of gait identification based on enhanced gait energy image （E-GEl） in informal angle of view is pro- posed to resolve the gait recognition in informal angle of view. The method of viewing angle transform unifies the angle of view. The background subtraction is adopted to extract the silhouette of human body. The gait cycle detecting method is in- troduced to confirm the gait cycle. The E-GEl is obtained according to the parameter model of human body skeleton. The 2-D PCA （principal component analysis） is employed to extract the eigenvector. The nearest threshold approach is used for classification. The experimental results show that the proposed E-GEl is better than the normal GEl in recognition effect.

关 键 词：步态识别 加强步态能量图 二维主分量分析 非规范视角 

分 类 号：TN919-34[电子电信—通信与信息系统] TP391[电子电信—信息与通信工程]