Optimization of EPZS algorithm based on motion estimation

 LI Jun1, ZHOU Xiao-hui2, YE Wen-long1
(1.College of Computer Science, Hangzhou Dianzi University, Hangzhou 310018, China;
2.School of Automation, Hangzhou Dianzi University, Hangzhou 310018, China)

Abstract: The real-time of encoder is reduced seriously because of the tremendous computation of the motion estimation, which has to cost 80% time at least in the whole encode period. In order to save the time of motion estimation and encode as far as possible, a method for optimizing the initial center point and search template was presented for the EPZS(enhanced predictive zonal search) algorithm of the H.264 video coding standards. Under the same case of image quality, the four testing sequence compared with the original algorithm was tested in the reference model JM10.1 of the H.264 encoder. The experimental results show that the performance of optimal EPZS algorithm is better and more strong adaptability, especially the effect of the more complex sequences is more obvious by testing the four testing sequence.
Key words: motion estimation; predictive vector; search template; enhanced predictive zonal search(EPZS)