Segmentation of Video Image Using Changing Detection and Block Based Approach

摘要

Segmentation of object from video image sequence is very important in many aspects of multimedia application. We present an approach to discover and segment foreground object(s) in video. Our approach is to identify 8x8 blocks in any frame and compute a series of binary partitions among those blocks and compare each frame of blocks finding change of all frames in images. Our approach is automatically segmentation of object on the persistent foreground regions of interest. We apply our method to challenge standard videos, and show competitive or better results than the state-of-the-art.