自适应联合相邻运动矢量的运动补偿插值算法

A Joint Motion-Compensated Interpolation Algorithm Using Adjacent Block Motion Vectors Adaptively

帧率上采样作为一种视频后处理技术,通过对原始视频插值得到高帧率视频,满足人们对高帧率视频的需求。传统基于重叠块的插值补偿算法会带来块效应或鬼影现象。为了解决这一问题,人们提出联合匹配块及其相邻块预测的方法,但是复杂度高,效果不明显。基于以上几点,本文提出一种自适应联合相邻运动矢量的运动补偿插值算法(Joint Motion-Compensated Interpolation Algorithm Using Adjacent Block Motion Vectors Adaptively,AJ-MCI)。在该算法中,将运动矢量矫正和运动补偿插值算法相结合,同时运动补偿模块中自适应地联合相邻匹配块,以最大限度刻画真实运动轨迹。实验结果表明,本文算法能很好的提升视频的主客观质量,同时保持较低计算复杂度。

As a video post-processing technique, frame rate up-conversion satisfies people＇s demand for high frame rate video by interpolating the original video. The traditional block-based interpolation compensation algorithm brings block or ghost image. In order to solve this problem, some people proposed a method of jointing matching block and its neighbor block prediction, but it has high complexity and little effect. Based on the above points, this paper presents a joint motion-compensated interpolation algorithm using adjacent block motion vectors adaptively. In this algorithm, the motion vector correction and motion compensation interpolation algorithm are combined, and the motion compensation module adaptively joins adjacent matching blocks to describe the real motion trajectory to the maximum extent. Experimental results show that the proposed algorithm can improve the subjective and objective quality of video, while maintaining a low computational complexity.