New Selective Block Matching Searching Algorithm Based on Block Discriminator Values For Motion Estimation

Abstract

In this paper, a new, fast and efficient selective block matching algorithm based on block discriminator values is suggested. The searching scheme consists of three stages: block discriminator values calculations, open full window then cascaded selection sub stages have been applied where each sub stage uses distinct block discriminator value to select sub set of candidate search points and finally perform block distortion measurement calculation to the latest candidate search points to find the best match. The famous Exhaustive Search (ES) and Three-Step Search (TSS) algorithms are used to compare with proposed algorithm. The experimental test shows that the proposed searching method results in compare with TSS algorithm, produce results little slower in term of search time but better in term of distortion measurement when the size of search area is [-6,6].Results in compare with ES show that the proposed searching method produces results near the ES in term of distortion measurement but faster than it in term of search time when the size of search area is [-7,7].