为了降低子像素精度运动估计的复杂度,提高运动估计的运行速度,基于泰勒定理以及泰勒展开式,提出一种不进行插值滤波的子像素精度运动估计方法。该算法以泰勒近似为基础,通过对率失真优化准则进行分析以及近似,以解二元一次方程组的形式获得子像素精度的运动矢量。实验表明,相对于进行插值的子像素运动估计方法,该方法能在保持运动补偿精度的基础上大幅度提高子像素精度运动估计的运行速度。
To reduce the complexity of sub pixel motion estimation and speed up the process of motion estimation,this paper proposed a novel sub pixel motion estimation algorithm without interpolation.This algorithm obtained the sub pixel motion vectors by solving a binary equation group which was from the Taylor approximation of rate-distortion optimization criterion.The simulations show that the proposal can speed up the sub pixel motion estimation significantly with keeping the accuracy of motion estimation compared with the sub pixel motion estimation with interpolation.