Fast motion estimation algorithm based on adaptive search range decision

Won Jin Lee, Jechang Jeong · 한국멀티미디어학회 국제학술대회 · 2011

The motion estimation in video codec is very important because motion estimation has decisive effect on encoding time and video quality. The most fundamental motion estimation method is Full Search algorithm. Although Full Search algorithm shows the best video quality, computation complexity is extremely increased. Thus, a variety of fast motion estimation algorithms have been proposed to reduce the complexity and maintain the high video quality. In this paper, we propose the new method that changes the search range using Sum of Absolute Difference between macroblock of current frame and macroblock at the same position of previous frame. According to experimental results, our proposed algorithm showed in the average of 0.03 ㏈ PSNR degradations and the computational complexity was 139 times less than FS.

Read the paper · More papers on PaperTik