An arbitrary shaped pattern selection algorithm for very low bit-rate video coding focusing on moving regions

Very low bit-rate video coding using patterns to represent moving regions in macroblocks has very good potential for improved coding efficiency. It is intuitively evident that instead of using a codebook of predefined patterns, superior coding performance is achievable if an approximation of shape is made by using a set of patterns dynamically extracted fiom the video content. This paper, proposes an Arbitrary Shaped Pattern Selection (ASPS) algorithm for very low bit rate coding. ASPS generates a specific pattern set of user-selectable size by applying a classifier, with the subsequent coding b e i i performed using these patterns. Experimental results show that ASPS exhibits better performance compared with other pattern matching algorithms, including the low-bit rate video coding standard H.263.