摘要
为了解决传统六边形搜索算法使用固定模板和固定的搜索步骤所带来的缺陷,提出了基于运动矢量中心偏置特性的改进算法。该算法充分考虑了运动矢量中心偏置这一特性,同时对传统六边形模板进行了改进。在保持六边形模板的优点的同时,使得其在搜索过程中更具有方向性。最后,对提出的算法进行了实验,结果表明,在保持图像质量的情况下,搜索速度有大幅的提高,证明了该算法的有效性。
To solve the flaw in fixed stripper and fixed searching step in traditional hexagon searching algorithm, an improved HEXBS algorithm based on center-biased characteristics is proposed. This algorithm takes full account of movement vector center-biased characteristics, and the traditional hexagonal template has been improved. While maintaining the advantages of hexagonal template, their search becomes more directional. Finally, the proposed algorithm has been experimented. The results show that the searching speed has largly increased in the case of a little drop in PSNR.
出处
《北京信息科技大学学报（自然科学版）》
2010年第1期70-73,共4页
Journal of Beijing Information Science and Technology University
基金
北京市教育委员会科技发展计划项目(KM200811232006)
关键词
运动估计
六边形
中心偏置
自适应
方向性
motion estimation
hexagon
center-biased
adaptive
directivity
