基于图像背景减除技术的步态识别研究  

Gait Recognition Based on Background Subtraction

作　　者：王纯贤[1] 吴杰[1] WANG Chun-xian, WU Jie （School of Machinery and Automobile Engineering, Hefei University of Technology, Hefei 230009, China）

机构地区：[1]合肥工业大学机械与汽车工程学院,安徽合肥230009

出　　处：《电脑知识与技术》2008年第7期120-123,共4页Computer Knowledge and Technology

摘　　要：步态识别是一种新的生物特征识别技术,旨在根据人们走路的姿势进行身份识别。本文利用图像背景减除技术进行步态轮廓检测,然后利用傅立叶描述子对步态轮廓图像进行描述,进行维数压缩,得到模板匹配,最后,利用最邻近法进行识别。实验证明,该算法具有较高的识别率。Gait recognition is a new technology of recognition based on biologic character, it aims to recognize the human＇s identity based on their style of walking. In this paper, we detect gait contour by using background subtraction, at the same time, we adopt Fourier descriptor to describe gait contour, using this method, we can reduce the dimension of data and create template. The recognition is achieved by nearest algorithm. Tests proved that this method have achieved comparative high correction identification ratio.

关 键 词：步态识别 图像背景减除 傅立叶描述子 最邻近法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术;自动化与计算机技术—计算机科学与技术]