计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (29): 164-167.

• 图形、图像、模式识别 • 上一篇    下一篇

基于时间保持能量图和Radon变换的步态识别

刘玉淑   

  1. 复旦大学 计算机科学技术学院,上海 200433
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-10-11 发布日期:2011-10-11

Radon transform-based chrono-gait image for gait recognition

LIU Yushu   

  1. School of Computer Science,Fudan University,Shanghai 200433,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-10-11 Published:2011-10-11

摘要: Radon变换把图像从坐标空间映射到Radon空间,因其可以保存频率信息而被应用在步态识别算法中。主要从频率角度入手,着力提高基于Radon变换的步态识别算法的识别正确率,提出了基于时间保持能量图的Radon变换步态识别算法。传统的步态能量图是对步态周期中经过归一化的人体轮廓图求算术平均而得到的步态特征表示,最近提出的时间保持能量图在保持步态能量图的优点的基础上,保留了步态序列的时间信息,在改进的步态周期检测算法的基础上,提出将时间保持能量图和Radon变换结合到一起的步态识别算法。也对结合不同数据空间的特征如频率、形状等做了初步探讨。

关键词: 频率, Radon变换, 步态能量图, 时间保持能量图, 步态识别

Abstract: Radon transform is an important method in gait recognition which captures the frequency information of an image by projecting the 2D image from I(x,y) space to Radon space.A Radon transform and chrono-gait image combined gait recognition algorithm is raised.Traditionally gait sequences are handled in a video with Gait Energy Image(GEI) which is a compact representation of gait features by averaging aligned and unified gait silhouettes.However it also loses temporal information.Chrono-gait images solve this shortcoming of GEI while sharing all the advantages of GEI.This paper effectively combines Radon transform with chrono-gait image to improve the performance.Experiments indicate the promising results of the proposed approach.The proposed method achieves high recognition accuracy.

Key words: frequency, Radon transform, Gait Energy Image(GEI), chrono-gait image, gait recognition