Complexity scalable motion estimation based on modes preselection in H.264

H.264 allows tree-structured partitioning motion estimation performing on multiple reference frames. This new feature improves the prediction accuracy of inter-coding blocks significantly, but it is extremely computational intensive. The complexity of motion estimation increases linearly with the number of used reference frames and is in proportion to the number of modes. This disables some applications on computations-constrained platform. Complexity scalable algorithm is an algorithm that has several computation scales, each of which adapts to certain computation power platform. In this paper, we propose a complexity scalable algorithm for motion estimation in H.264 based on mode preselection in multiple reference frames. This proposed algorithm has 4 computation scales consisting of 72%, 47%, 33% and 20%. Simulation results show that this algorithm can effectively reduce complexity of motion estimation with gracefully degraded quality from the original.