An efficient technique for segmentation of key object(s) from video shots

The most fundamental task in video processing is to partition long video sequences into a number of shots and find a key frame of each shot for indexing and browsing. In this paper, we want to extract key object(s) instead of key frame. In order to do that, we extend our previous technique for shot boundary detection (SBD) to select two candidate frames out of each shot, then propose a new framework to segment object(s) from those selected frames using color quantization and background adjustment. We make our scheme cost-effective and automatic by avoiding expensive computations, and removing manual processing. Experimental results show that the proposed scheme is promising.