An improved search algorithm for motion estimation using adaptive search order

An adaptive search order (ASO) algorithm is presented to speed up the block motion estimation in digital video coding. According to the motion trend, a table of the adaptive search order is defined. For each searching iteration, a better search order is derived and then the best matched block can be found in the early search stage. Some experimental results demonstrate the computational advantage of the proposed improved algorithm when compared to previous algorithms, such as the full search algorithm, the successive elimination algorithm, the block sum pyramid algorithm, and the multilevel successive elimination algorithm.