基于区域特征的步态识别研究

Study of Gait Recognition Based on Region Feature

步态识别主要是以人走路的姿势来识别其身份。为了能准确快速地识别,本文依据步态的周期性,以一个步态序列中的6帧特殊图像来描述步态的变化,再提取步态髋关节以下部分作为研究对象,确定髋关节以下部分的质心,以质心为原点,建立坐标系。将髋关节以下的部分划分成36个小区域,以周期变化的区域面积作为特征,描述步态的周期变化。这样既可以较全面地描述一个步态的信息,又降低了维数,减小计算量。同时引入最近邻分类器进行分类。实验证明,该算法不仅能得到较好的识别率,并且计算速率快。

Human gait recognition is the process of identifying individuals by their walking manners.In order to identify quidkly and accurately,we determined a gait sequence of the six specific graphics based on the cyclical nature of gait.Thus,the lower part of the hip was extracted as the research object,and centroid of the object was determined.Then, a reference frame was built with the centroid as the origin.The areas of coordinates were divided into 36 small part,the periodic changes in size of the area was extracted as the features to describe the gait cycle.In this way,we can retaint the most useful features and reduce the number of dimensions.In this paper the NN classifier is also used,experiments showed that it not only can get a better recognition rate,and calculation rate is enhanced quickly.