摘要
提出了一种新的快速、有效的块匹配运动估计算法。算法采用多步搜索方法 ,利用相邻块之间的运动相关性 ,选择反映当前块运动趋势的预测点作为初始搜索点 ;用子采样块匹配失真度量来减少计算量 ;利用运动矢量的中心倾向的分布特性 ,用多侯选点钻石形状搜索方法来提高运动估计的速度和准确性。实验表明 ,该算法计算复杂性低 ,预测质量好。
Block motion estimation using full search is computationally intensive,many fast algorithm have proposed to reduce the computation at the expense of less accuracy of motion estimation.In this paper,we present a new fast and efficient search algorithm for block motion estimation. The proposed algorithm is based on the ideas of predicted starting search point,subsampled block distortion measure,center\|biased distribution of motion vector,multiple\|candidate diamond search.From the experimental results,the proposed algorithm is superior to many other well\|known fast algorithms in both quality performance and computational complexity.
出处
《北京大学学报（自然科学版）》
CAS
CSCD
北大核心
2001年第3期315-322,共8页
Acta Scientiarum Naturalium Universitatis Pekinensis
