Priority-Based Normalized Partial Distortion Search Algorithm for Fast Motion Estimation

In this paper, a priority-based normalized partial distortion search algorithm for fast motion estimation is proposed. The proposed algorithm is mainly composed of the two algorithms, the priority-based normalized partial distortion and the dynamic search range decision algorithm. Firstly, the distortion behavior of each macroblock is analyzed in this paper to point out the calculation priority/order of each pixel. For the partial distortion early termination, the normalized partial distortion search algorithm is applied to half stop the error calculation. In order to further increase the coding speed, a dynamic search range decision algorithm is used for changing the search range automatically. Through the proposed algorithms, the significant computational complexity reduction can be achieved while maintaining the ignorable PSNR degradation