Human Detection in Video Based on Template Non-parametric Statistical Framework
基于模板的非参量统计的视频图像中人的检测跟踪方法

In order to detect human in images under lower resolution,complex environment and transformation of human self were overcome.The body of human was divided to several parts which are represented as linear model of color for overcoming these problems.The feature of body could be robust through the composition of these linear models.Based on Mean shift,a template using relative position was combined with Mean shift to get a detecting method which is better than single Mean shift.