A fast full search algorithm using directional oriented search order

Abstract

We propose a fast full search algorithm that reduces the computation of block matching algorithm which is used for motion estimation in the video coding. The conventional spiral search algorithm is good for slow motion picture because it starts searching at the center of the search region and moves the search point pixel-bypixel. The proposed algorithm, however, is the efficient motion estimation method which is good for fast and slow motion picture. It estimates the motion in the new search order after partitioning the search window and adapts the directional search order in the re-divided search window. At the result, we can check that the proposed algorithm, which is based on MSEA, reduces the computation by up to 99% for block matching error, compared with the conventional full search algorithm, without any loss of image quality.