Motion estimation based on spatio-temporal correlations and pixel decimation

Abstract

Based on motion estimation using spatio-temporal correlations, an enhanced search algorithm is proposed to further improve the performance in terms of reducing total amount of computations while preserving the video quality. Our experimental results show that the proposed algorithm has substantially smaller computational complexity having some quality reduction. For sequences with low motion PSNR reduction is around 0.1 dB and computational complexity is around 1700 times less than that of a full search. © 2005 SPIE and IS&T.