Kernel Density Estimation Method Basing on Color and Motion Features Frame for Moving Object Detection
Yu Ting Guo, Zi-Qiang Shen · 2015
Due to large calculations and complex background updating problems of kernel density estimation, this paper proposes a feature frame building method based on the combination of color feature and motion information, using this method to extract the number of samples N, it can not only reflect the global information of image but also reflect local information variations of image, besides it can effectively solve the inaccurate problem of the sample numbers, thereby enhancing the instantaneity of kernel density estimation algorithm. Keywords-kernel