一种低编码复杂度可分级的视频编码系统

A scalable video coding system with low encoding complexity

在可分级视频编码(SVC,scalable video coding)的框架下,利用分布式视频编码(DVC,distributed video coding)技术,设计了一种低编码复杂度的SVC方案。该系统具有空间可分级的特性,各分层中仅用到了传统的帧内编码技术和DVC技术,最大限度的减小了SVC系统的编码复杂度。在该系统中,充分利用分级系统的特点,在增强层(EL)的解码中提出了一种基于二次搜索和残差补偿(DSRCB)的边信息(SI)生成算法和一种基于时空域虚拟噪声模型的估计算法,并针对各分层图像的频域特性优化了量化模型。实验表明,与基于传统视频编码技术的SVC系统相比,该系统具有极低的复杂度,性能超过了非可分级的DVC系统,而且在较小的GOP(group of pictures)尺寸下获得了接近传统SVC系统的性能。

A scalable video coding（SVC） structure with low encoding complexity is designed based on the distributed video coding（DVC） technology.The structure is spatially scalable.Only conventional intra coding technology and DVC technology are used in all coding layers so as to keep the encoding complexity in a much lower level.A double search and residual compensation based（DSRCB） side information（SI） generating algorithm and a temporal-spatial based（TSB） virtual noise model are proposed for the enhancement layer（EL） video decoding,taking fully advantages of the scalability in the system.According to different characteristics of the base layer（BL） and the EL frames in scalable video,the quantization models in both layers are also optimized.Simulations show that the designed SVC system has much lower encoding complexity and similar performance compared with a conventional SVC system with the same structure of scalability.