改进的EPZS运动估计算法研究

ON IMPROVED MOTION ESTIMATION ALGORIT HM OF EPZS

运动估计是视频压缩编码中最重要和最耗时的环节之一,研究高效快速运动估计算法是视频压缩编码中的核心问题之一。H.264编码器由于采用了大量高精度的运动矢量,计算量和估计消耗时间都迅速增长。针对被JM模型采纳的EPZS算法进行了多层次的匹配策略和基于先验信息的自适应阈值技术的优化,有效地降低了算法运动估计的复杂度和收敛速度。实验结果表明,优化后的算法与原算法相比,在重建图像质量和码率接近的情况下,可以节省12%的运动估计时间和7%的编码时间,提高了编码器的实时性。

Motion estimation is one of the most significant a nd time-consuming parts in video compression encoding area.To find out the fast and efficient motion estimation algorithm is one of the core issues in video com pression encoding.Due to the use of a great deal of motion vectors with high pr ecision,the complexities of computation and estimated time of the H.264 encoder increase dramatically.Aiming at the EPZS algorithm adopted by the JM model,th is paper proposes an optimised algorithm from two aspects,including a condensed hierarchical block matching scheme and a priori information-based adaptive thr eshold technology,which effectively reduces the complexity of the algorithm in motion estimation and the convergence speed.The results of experiment proves th at,compared with the original algorithm,the optimised algorithm can save 7% ti me in encoding and 12% time in motion estimation,while achieving almost the sa me PSNR and the same quality of the reconstructed video image.