An improved block matching algorithm based on successive refinement of motion vector candidates

Abstract

This paper describes a new block matching algorithm based on successive refinement of motion vector candidates. The proposed algorithm starts with a full search where an approximated matching criterion is used to obtain a set of candidates for motion vector. In each successive searching process, the matching criterion becomes refined and the search is performed only for the candidate set obtained at the preceding layer to refine the candidates. By repeating this process, at the last layer, only a single motion vector can be selected from a few candidates using the conventional mean absolute difference criterion without approximation. Since a full search is performed at the first layer, even though a coarse matching critirion is used, the proposed algorithm diminishes the local minimum problem in the existing fast search algorithms and also reduces the computation time drastically compared with the full search algorithm.