摘要：Region based video coding has been a very active research area over the last few years. It has been viewed as a potential an alternative to traditional schemes suffering from the “blockiness” at very low bit rates. The success of region based techniques relies on the ability to describe regions in an image succinctly by their shape and sizes. In this paper we present a new approach to region based image coding. It is based by incorporation the knowledge of the border in image segmentation and exploits temporal redundancies between successive images
