摘要
提出了一种基于在整像素搜索结果的± 1像素邻域内 SAD函数随着与亚像素全局最小点距离的增加而单调增大的半像素两步快速搜索算法。本算法相继在水平和垂直两个共轭的方向上搜索半像素 SAD值的最小点 ,从而得到运动矢量的半像素分量。同时还采用了在每次计算 SAD值之前按需对块进行插值的方法。仿真结果表明本算法在保持图像质量和码流长度基本不变的同时使运算量下降到原来的一半以下。
A fast search algorithm for half-pixel motion estimation is proposed based on the characteristic. The SAD distortion function increases monotonously as the search location moves away from the global sub-pixel minimum within ±1 pixel neighborhood of the result of integer-pixel search. The algorithm sequently searches the minimum SAD within the neighborhood on the horizontal and vertical directions so as to find the half-pixel component of the motion vector. At the same time, the necessary half-pixels are only interpolated whenever computing SAD . Simulation results show that the algorithm is simple and efficient, and requires less than 1/2 in the computation for the full half-pixel search while keeping the same image quality and bit rate.
出处
《数据采集与处理》
CSCD
2004年第1期51-55,共5页
Journal of Data Acquisition and Processing
