S. Jin, S.-J. Park, and J. Jeong (Korea)
Fast Full Search, Adaptive Region Order, Adaptive Calculation Order, and Fast Motion Estimation
In this paper, an adaptive fast full search algorithm is proposed. In the first of two proposed stages, using a predicted motion vector, we determine the adaptive region order based on gradient distances among all regions after partitioning the search range. In the second stage, by means of adaptive region order, the matching error of all search positions is calculated in an optimized calculation order which is derived from the local image complexity of the current macro-block to search. Based on a multilevel successive elimination algorithm the proposed algorithm reduces the complexity of motion estimation up to 96.1% on average, compared with the conventional full search algorithm without any loss of image quality.
Important Links:
Go Back
