Fusion side information based on genetic algorithm in distributed multi-view video coding

Generating side information (SI) utilizing intra-view correlation and inter-view correlation efficiently is one of the key technologies in distributed multi-view video coding (DMVC). Generally, different SI got from different generation method is applicable specifically to different motion situation. And the fusion SI generated by the fusion technology may be hopefully applicable to most of motion situations. In this paper, we present a new method to get fusion SI with the help of genetic algorithm (GA) to mix the advantages of the inter-view SIs and the intra-view SIs. Experimental results show that up to 1.2dB of PSNR improvement can be obtained for SI generation with our method.