An Adaptive and Efficient Selective Multiple Reference Frames Motion Estimation for H.264 Video Coding

Abstract

In the popular video coding standard H.264/AVC, many advanced techniques are employed. One important technique is the use of multiple reference frames motion estimation. However, the computational load increases with the number of references frames. In this paper, we suggest a selective multiple reference frames motion estimation (SMRFME) architecture which takes use of the information of the 1st reference frame to determine whether it is necessary to search remaining reference frames. In addition, three early termination schemes are applied to the remaining reference frames of the candidate modes. The simulation results demonstrate that the proposed algorithm can achieve up to 77% of time saving compared to the multiple reference frames full search algorithm, while maintaining a high coding performance.KeywordsH.264multiple reference framesselective multiple reference frames motion estimation (SMRFME)early terminationAZBregion based