A New Approach to Fast Multiple Reference Frame Motion Estimation for H.264

Abstract

In the novel video coding standard H.264/AVC, motion estimation is allowed to adopt multiple reference frame in order to improve the coding efficiency. However, the computational cost of multiple reference frame motion estimation is very high. In this paper, we proposed a fast multiple reference frame motion estimation algorithm which takes into account the correlation and center-biased characteristic of motion vector in multiple reference frame adequately. The method makes a selection for multiple reference frame in advance, so it can improve the coding efficiency greatly. Experimental results show that the proposed method can reduce encoding time by 30.91% on average with less change of rate distortion performance.