摘要
步态识别主要通过人体走路的姿势来识别人的身份.从经过背景减除后的人体运动图像序列中,根据面积变化确定运动周期,提取人体宽度信息,对下肢进行Radon变换,经处理后提取运动角度信息.对所得到的信息主分量分析变换后进行动态时间规整,采用最近邻分类器分类.该方法可以有效地降低人体运动时身体自遮挡及影子带来的影响.在小数据库上取得了很高的识别率.实验结果表明该方法有效.
Human gait recognition is the process of identifying individuals by their walking manners. The gait as one of biometrics has recently gained more and more interests from computer vision researchers. In this paper, the objects are got through subtracting backgrounds from the image sequences. A gait cycle is obtained through analyzing the changes of the objects' areas. The width features are extracted from human body. Radon transform is used to analyze the low limps and the limp angle features are got. After principal component analysis (PCA) transforming, dynamic time warping (DTW) is used to obtain the distances be- tween the probe and the training sequences. The nearest-neighbor classifier is used to distinguish the classes. The experimental results demonstrate that the method is effective.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
北大核心
2007年第3期301-304,共4页
Journal of Harbin Engineering University
关键词
步态识别
背景减除
RADON变换
主成分分析
动态时间归整
gait recognition
background subtraction
Radon transform
principal component analysis
dynamic time warping