H.264 Interframe Coding Mode Decision Based on Classification and Spatial Correlation

Abstract

To select the best model for the current macroblock, the encoder need to traverse all the possible models, each model calculated the cost of coding needed, select from the best model for minimum cost, this method is called full analysis. Full analysis of course be able to choose the best coding mode, it is clear that computation is enormous, according to experimental statistics, H.264/AVC motion estimation and compensation is calculated in the standard module of the largest, accounting for the entire Encoder 1/3, and mode decision algorithm as a core, a direct impact on the efficiency of motion estimation and compensation. It is noteworthy that the decision based on rate-distortion model and the model based on absolute error and the decision there are some differences, because the decision of this thesis is concerned with modeling algorithms, and therefore to the discussion of simplicity, this all ideas will be discussed and algorithms are based on the absolute error and the standard model of decision.