论文部分内容阅读
针对分布式视频编码(DVC)系统中固定周期关键帧选取(PKFS)方法忽视了帧间相关性的缺陷,提出了一种自适应关键帧选取(AKFS)算法。利用图像特征点检测与匹配的方法,将相邻图像的非匹配点作为帧间相关性的近似,把累积或平均非匹配点数超过阈值的帧判定为关键帧。在此基础上,提出改进的帧内插方案,以适应不同长度序列组的边信息生成;将零运动强度的关联帧合并为一帧图像参与编解码,进一步提高了系统的压缩效率。实验结果表明,对于不同运动特性的序列,本文提出的算法可以明显提升边信息帧的重建质量,使系统的率失真性能提高0.9～2.0 dB,并有效降低了编码传输码率。
Aiming at the defect that the fixed periodic key frame selection (PKFS) in DVC system ignores the inter-frame correlation, an adaptive key frame selection (AKFS) algorithm is proposed. Using the method of image feature point detection and matching, the non-matching point of the adjacent image is taken as the approximation of the inter-frame correlation, and the frame whose accumulated or average number of non-matching points exceeds the threshold is judged as the key frame. On this basis, an improved frame interpolation scheme is proposed to adapt to the generation of edge information of different length sequence groups. The correlation frames of zero motion intensity are combined into one frame of image to participate in coding and decoding, which further improves the compression efficiency of the system. The experimental results show that the algorithm proposed in this paper can significantly improve the reconstruction quality of edge information frames and improve the rate-distortion performance of the system by 0.9-2.0 dB for sequences with different motion characteristics, and effectively reduce the coding bit rate.