Region-based moving object detection Using SSIM

In this paper, a region-based moving object detection based on SSIM is concerned. SSIM is commonly used to measure the similarity of two images. This paper first analyzes the characteristics of SSIM value. Then the possibility of moving target detection using SSIM is analyzed. Finally the single Gaussian model based on the SSIM values of area is used to detect moving object, and compare the result with traditional single Gaussian model based on gray level. The experimental results show that the proposed method is better than traditional single Gaussian model based on gray level, and the detected moving object has no empty holes in side. The shadow is suppressed to a certain degree.