基于三维步态特征的步态识别方法

Design and Research on 3D Gait Recognition System

研究步态识别问题,针对在当前二维步态识别系统中,识别过程仅仅针对灰度、平面几何距离等二维特征信息,忽略了人体走路时的三维步态特征,步态识别准确度不高的问题。提出了一种加入三维参数的步态识别算法。利用摄像机采集单帧步态图像序列,利用身体结构的知识和摄像机标定的知识提取出人体走步时的人体三维特征数据,利用提取出二维和三维的步态特征,进行步态识别。结果表明相对于以二维步态特征为参数的步态识别,识别率有了明显改进。

Researching gait recognition.Identification process is only aiming at the two-dimensional feature information of gray and plane geometry distance,and ignores the three-dimensional gait characteristics of human walking,so that the gait recognition accuracy is not high.To solve this problem,this paper presents a gait recognition algorithm with three dimensional parameters.The knowledges of physical structure and camera calibration are used to draw the three-dimensional data when human walking,and the characteristics of two-dimensional and three-dimensional gait are used for gait recognition.The results show that compared with two-dimensional gait recognition,the method can improve the recognition rate significantly.