快速运动估计与AGS结合的可伸缩性视频编解码系统

可伸缩性视频编解码(SVC)系统是一种能够实现时间、空间、质量等可调节可伸缩的视频编解码技术。它是通过采用运动补偿时域滤波(MCTF)和运动估计(ME)技术结合有损编码及熵编码等技术,实现对视频序列进行成组编解码的任务。其中,如何更加有效地提高对时间相关性的利用以及如何根据视频性质进行自适应调节图像组(GOP)大小问题是提高系统编码效率的关键环节。文中通过采用一种利用时间信息的快速运动估计技术及对应的自适应GOP结构(AGS),有效提高了可伸缩性视频编解码系统的编码能力。

Scalable video coding-decoding system based on the combination of fast motion estimation and AGS

Scalable video coding (SVC) system is a video coding system which can realize scalability in aspects such as temperal field, spatial filed and coding quality. By combining motion compensated temporal filtering (MCTF) and motion estimation(ME)methods with lossy coding and lossless entropy coding techniques, the task of video coding in group is achieved. Therefore, the effective exploitation of the temporal information and the realization of an adaptive group of picture (GOP) structure based on the motion activity of the coded sequence are important factors to improve the performance of SVC system. A fast motion estimation algorithm based on robust temporal information and the corresponding adaptive GOP structure (AGS) are employed. The result shows that the coding performance of SVC system is enhanced obviously.