针对H.264改进的快速整像素运动估计算法

Improvement of fast integer pixel motion estimation algorithm for H.264

以视频压缩标准H.264联合开发模型(JM)中的运动估计算法UMHexagonS为基础,提出了一个新的快速整像素运动估计算法来改进压缩编码性能。在起始搜索点预测部分,提出了新的预测运动矢量(MV)检测顺序,以提高起始搜索点的准确度;在全局搜索部分提出了自适应全局搜索方法,根据准确度最高的两个预测MV之间的关系,适当跳过非对称十字型模板搜索和非均匀多重六边形模板搜索,并通过对不同序列的测试,验证了判断准则的可行性与准确性。根据实际序列中最佳MV相对起始点的分布,提出了改进5×5搜索。另外,增加了针对子宏块的提前终止策略,在不增加额外运算量的前提下,进一步减少了运动估计开销。实验结果表明,相对UMHexagonS算法,提出的改进算法使搜索点总数平均减小了83.80%,信噪比平均下降了0.021dB,或输出码率等效增加了0.46%。该算法有效降低了运动估计的运算量,而只带来了很小的编码性能下降,且对不同运动强度的视频序列具有均匀的算法效果。

On the basis of the UMHexagonS algorithm used in the Joint Model（JM） for H. 264, a new fast algorithm on integer pixel motion estimation was proposed to improve the video encoding charac- teristics. A new order of checking predicted Motion Vector（MV） was proposed in the predicting start- ing search point to promote the accuracy of starting search point. Then, a self-adaptive global search method was proposed for the global search, by multi-hexagon-grid search could be skipped based highest accuracy, and the feasibility and accuracy a number of different sequences. Meanwhile, an which the unsymmetrical-cross search and uneven on the relationship between two predicted MVs with of this judging method was verified through tests on improved 5 × 5 search method was presented according to the distribution of best MV relative to the starting search point and a new early termination technique for sub-macroblock was added to further reduce the cost of motion estimation without addi- tional computation. According to the experiment results, the proposed algorithm reduces the total number of search points by 83.80% and the Peak Signal to Noise Ratio（PSNR） about 0.021 dB on av- erage and increases an average increment of 0.46% on equivalent bitrates as compared with that of o- riginal UMHexagonS algorithm. With a negligible performance degradation, the proposed algorithm reduces the computation of motion estimation effectively and offers a well-distributed effect on se- quences with different motion intensities.