A new motion estimation algorithm

Abstract

The block-matching algorithm (BMA) is widely used for image sequence coding. In the conventional BMA, the blocks within a frame are processed independently and the intensity variations are not considered. The authors present a motion estimation algorithm in which the correlation of interblock is used to save the computations, and the variations of image intensity are considered for the improvement of estimated motion vector accuracy. The simulation results show that the performance of the motion estimation algorithm is better than that of the 2-D logarithmic search method.<>