基于时空预测向量相关性的运动估计算法

针对UMHexagonS算法体现出来的问题，利用时间预测向量和空间预测向量的位置映射关系，提出了一种新的运动估计算法--基于时空预测向量相关性的运动估计算法。该算法首先在小范围得到最优点后，继续利用预测矢量的时空方向相关性进行特定方向的扩展搜索，避免了提前落入局部最优点，并减少了搜索点数，从而提高了搜索质量。实验结果表明，与UMHexagonS算法相比，该算法在保持码率基本不变的情况下，能有效地减少运动估计时间，并且能一定程度地提高单帧的峰值信噪比。关键词：

A motion estimation algorithm based on the correlation of spatial-temporal prediction vector

Using the mapping relationship of the temporal prediction vector and the spatial prediction vector, a new motion estimation algorithm is proposed to solve the shortcoming of UMHexagonS algorithm. In order to avoid the early fall into the local advantages, reduce the search points and improve search quality, it gets the most advantage of the algorithm on a small scale and continues to expand the search for a specific direction of using the directional prediction vector. The experimental results show that, compared with UMHexagonS algorithm, the new algorithm can generally improve the single frame peak signal to noise ratio and effectively reduce the time of motion estimation in case of almost the same bit rate.