Fast mode decision for P-Slices in H.264/AVC based on probabilistic learning

In this paper, a fast mode decision algorithm for P-Slices in H.264/AVC is proposed. The algorithm is based on probabilistic learning. Spatial correlation between neighboring macroblocks is exploited by a Bayesian classifier to predict the best candidate mode for the current macroblock. Early termination strategy is also designed for skipping unnecessary mode checking. Extensive experiments show that the proposed algorithm can achieve 60%-70% encoding time reduction compared to JM10.0, and only yields a negligible PSNR loss and little increment in total bits and bit rates.