论文部分内容阅读
H.264/SVC(Scalable Video Coding)使用了多种预测编码机制,极大地提高了分级编码的编码性能,同时也引入了复杂的编码依赖关系,给编码器的优化控制带来了极大的难度。因此,有必要研究这种不可忽视的编码依赖关系。在介绍H.264/SVC的主要预测编码机制的基础上,分析归纳出H.264/SVC的编码依赖关系主要来源于帧间预测、层间帧内预测和层间残差预测,然后针对这三种预测机制的特点,结合文献[1]提出的帧间依赖关系,基于实验数据,建立了编码依赖模型,得到了用编码依赖测度表达的编码依赖关系。最后,将时域依赖模型应用到H.264/SVC基本层码率控制中,实验数据表明时域依赖模型能够提高码率控制的性能。
H.264 / SVC (Scalable Video Coding) uses a variety of predictive coding mechanism, which greatly improves the coding performance of hierarchical coding, but also introduces a complex coding dependencies, to the encoder optimal control has brought a tremendous Difficulty. Therefore, it is necessary to study this can not be ignored coding dependencies. Based on the introduction of the main predictive coding mechanism of H.264 / SVC, it is concluded that the coding dependency of H.264 / SVC mainly comes from inter-frame prediction, inter-layer prediction and inter-layer residual prediction. Based on the characteristics of the three prediction mechanisms and the dependencies between frames proposed in [1], a code-dependent model is established based on the experimental data and the coding dependency expressed by the code-dependent measure is obtained. Finally, the time-domain dependent model is applied to the H.264 / SVC base layer rate control. Experimental data show that the time-domain dependent model can improve the performance of rate control.