A study on adaptive motion estimation algorithm using spatial and temporal correlation

Abstract

Block motion estimation using full search is computationally intensive. Previously proposed fast search algorithms limit the size of search window, the size of sequence (usually QCIF format) and search locations. These algorithms do not fit for a sequence having large motion or larger picture size. In this thesis, we present a new fast and efficient search algorithm for block motion estimation that produce better quality performance than TSS and 4SS and less computational time. The proposed algorithm is based on the idea of selecting an adaptive search window per macro-block using 3 spatial macro-blocks and one temporal macro-block with the distance criterion without computational burden. Simulation results show not only better performance in picture quality than that of TSS and 4SS, but also shorter execution time. Therefore it can be applied to the real time system which has limit in using resources.