基于运动连续性的帧率上变换算法  

Frame Rate Up-conversion Algorithm Based on Motion Continuity

作　　者：李珂[1] 高志勇[1] 陈立[1] 

机构地区：[1]上海交通大学图像通信与信息处理上海市重点实验室,上海200240

出　　处：《电视技术》2013年第11期61-65,共5页Video Engineering

摘　　要：为了解决高清视频序列固有的运动模糊、残影等问题,同时提高低比特率通信的质量,提出了一种新的帧率上变换算法。该算法将遗传算法引入到运动估计中并结合现有运动估计算法进行了改进,使其充分利用视频在时间上和空间上的连续性,提高算法性能,同时减小算法的复杂度;对得到的运动矢量场进行了矢量细化处理,减小了块效应;并利用邻域运动矢量进行加权的重叠块运动补偿,提高内插块的精度。实验结果表明,与传统算法相比,该算法在主客观评估中都有较大的性能提升。In order to solve the inherent drawback that most HD videos have motion blur and shadow on liquid crystal displays（LCDs）, and to enhance the quality of the video in low-bit communication at the same time, a novel Frame Rate up-Conversion algorithm is proposed. To fully take advantage of the spatial and temporal continuity, combines existing algorithms with Genetic Algorithm used in motion estimate, from which can improve the performance of the algorithms and reduce the computational complexity. It use the motion refinement to process the motion vector field, which is an applied procedure to reduce artifacts. It also adopts the weighted Overlapped Block Motion Compensation（OBMC）using neighborhood motion vectors to enhance the quality of image. Experimental resuhs show that the proposed algorithm achieves a better performance in subjective and objective assessments than the conventional algorithm.

关 键 词：帧率上变换 空间和时间连续性 遗传算法 重叠块运动补偿 

分 类 号：TN911.73[电子电信—通信与信息系统]