快速自适应运动估计算法

Fast self-adaptive motion estimation algorithm

为降低H.264中运动估计的复杂性,通过对自然图像序列的分析,提出了一种基于块的运动特征的快速自适应搜索算法。该算法充分利用视频序列中当前块的运动状态,根据不同块的运动情况合理选择运动搜索模板。实验表明,该算法在保证重构图像质量的前提下,编码速度有了显著提高,保证了实时应用的要求。

In order to reduce the complexity of motion estimation in H.264,through the analysis of natural graphic sequence,a fast self-adaptive searching algorithm based on the motion characteristics of blocks was proposed.The algorithm takes advantage of the motion status of the current block in video sequences and selects the proper motion search model according to the motion situations of different blocks.The experiment shows that,on the premise of the quality of reconstructive graphics,the proposed algorithm can speed up encoding significantly as well as meet the needs of real-time applications.