基于块金字塔的快速块匹配算法

在运动图象编码中，块匹配技术有着非常广泛的应用。在众多的匹配算法中，金字塔算法(blocks sum pyramid algorithm，BSPA)由于速度较快、算法比较简单而得到了广泛的应用。在金字塔算法的基础上，本文提出一种新的基于块金字塔的块匹配算法(NBSPA)。算法采用了预测初始中心，行更新等技术，实验结果表明，新的算法有效地降低了运算量。

A Fast Block Matching Algorithm Based on Block Sum Pyramid

Block matching technique has been widely used in video compression. A lot of algorithms have been proposed to accelerate the estimation speed, among those algorithms, block sum pyramid algorithm (BSPA) is widely used. In this correspondence, a new block sum pyramid algorithm (NBPSA) to motion estimation is presented. Compared with BSPA, NBSPA estimate the vector of the minimum mean absolute difference (MADmin). In the mean time, instead of the value updated level by level, the estimation of MAD is updated row by row from top to bottom. Experimental results showed that compared with exhaustive search approach in BSPA, NBSPA algorithm can be used to arrive at the same result with much less computational complexity.