To reduce the complexity of AVS sub-pixel motion vector search, this paper proposes a new improved algorithm based on little diamond window searching policies according to the law of the general video sequence motion vector focusing on the near of the initial searching point at the time of the subpixel motion estimation on the basis of deeping analysis to HFPS algorithm. Comparing with the sub-pixel HFPS algorithm search algorithm, the time of sub-pixel motion estimation can reduce 51.97％ on average and efficiently decrease the computational number of sub-pixel motion estimation on average PSNR lost less than 0.01dB to the video sequences with different motion characteristics.
Adaptive Interleaved Motion Vector Coding using Motion Characteristics