Fuzzy Based Optimum Bit Allocation for Rate Control in Video Compression

Abstract

This paper presents an algorithm for content-based bit rate allocation for video coding in order to improve ratedistortion (R-D) performance. A fuzzy tool is developed to analyze different areas of a frame. This tool attributes a level of interest to each macroblock. It is effective for a very large range of natural pictures. Beginning with the segmentation of the original picture, a non-hierarchical fuzzy model allows combining the results of distinct criteria in order to increase the reliability of the system and provide convergence to a meaningful final result. R-D optimal rate allocation is done taking help of the fuzzy model and human visual system. The model developed is incorporated in TM5 rate control algorithm and the results are compared with the same.