论文部分内容阅读
分数像素精确运动估计的改进是整个运动估计模块优化的关键,本文提出了基于H.264的内容自适应分数像素运动估计算法。首先,提出基于平坦区域宏块预测的无效分数像素运动矢量(MV)搜索省略算法(SMBP);然后,改进H.264采用的基于中心的分数像素搜索算法(CBFPS),提出基于预测矢量的增强型菱形模板(EDSP)搜索算法。实验结果表明,内容自适应分数像素运动估计算法比分数像素全搜索算法(FFPS)在峰值信噪比(PSNR)有微小降低(0.095～0.209 dB)的情况下,平均减少了75.6%的分数像素搜索点,整个运动估计模块平均节省了38.5%的计算量。
The improvement of fractional pixel precise motion estimation is the key to the optimization of the whole motion estimation module. In this paper, a content adaptive fractional pixel motion estimation algorithm based on H.264 is proposed. Firstly, an algorithm of fractional missing pixel motion vector (SMBP) search based on flat region macroblock prediction is proposed. Then, based on the center-based fractional pixel search algorithm (CBFPS) proposed by H.264, the predictive vector enhancement Type diamond template (EDSP) search algorithm. The experimental results show that the content-adaptive fractional pixel motion estimation algorithm has an average reduction of 75.6% of fractional pixels with a small reduction of peak signal-to-noise ratio (PSNR) of 0.095 ~ 0.209 dB Search points, the entire motion estimation module saves an average of 38.5% of the computation.