Novel hierarchical search motion estimation algorithm for mobile video transmission

Motion estimation is an important technology for mobile video communication. This paper presents a new large-scale sampling hierarchical search motion estimation algorithm which takes into account data fetching and memory management factors as well as the computation complexity for motion estimation implementation. The proposed scheme is suitable for real-time video communication with low computational complexity, reduced data fetching and simple memory access. The experimental results indicate the average decoding PSNR with the algorithm is only about 0.15 dB lower than that with the optimal full search scheme