中文摘要：

步态识别作为一种新的生物识别技术，通过人走路的姿势实现对个人身份的识别和认证。步态特征提取是步态识别的关键步骤。采用背景消减法与对称差分法相结合对运动人体分割，采用改进的GVFSnake模型对人体运动步态轮廓进行边缘提取。实验结果表明该方法能准确高效地提取边缘特征作为步态识别的特征。

英文摘要：

Gait recognition is a new biometric identification technology.It achieves personal identification and authentication through the way of people walking.The feature extraction of gait is the key step in gait recognition.This paper combines the background subtraction method withsymmetric differential method to segment the motion human image, and then extracts the contour of motion human with modified GVF Snake model.The experimental results show that the proposed method can extract contour as the gait recognition character effectively.