Skip to Main Content
Exhaustively searching over all possible partition modes using rate distortion comparison in H.264/AVC motion estimation to choose the best coding mode for each macroblock has much better coding efficiency than other existing standards. However, exhaustive searching has high computational complexity. This study analyses the statistics of coding mode distribution to determine the need of motion estimation for a current macroblock according to its previous corresponding macroblock information. A fast algorithm based on fuzzy reasoning with temporal and texture correlation of coding block information is proposed to reduce the computation time effectively. The fuzzy reasoning technique is used to determine the most appropriate mode quickly for motion estimation in H.264/AVC video coding. Simulation results show that the proposed algorithm has an encoding time 55.7-70- less than that of taken by searching for all coding modes exhaustively, while retaining the almost same coded picture quality and bit rate.