一种基于漏预测技术的FGS视频编码框架

FGS video coding framework based on leaky prediction techniques

精细粒度可伸缩性(fine granularity scalability,FGS)视频编码是MPEG-4标准的视频流化框架中的关键技术。以FGS编码框架为基础,针对FGS编码效率较低的特点,提出一种基于漏预测技术的视频编码框架。在改进的FGS编码框架中,通过引入漏预测因子,利用增强层参考和基本层的重构图像自适应形成增强层的时域预测信息,从而达到控制误差漂移,提高FGS的编码效率的目的。实验证明,改进的编码框架编码效率比FGS在同等条件下提高大约4dB。

Fine granularity scalability （FGS） video coding is one of the key techniques in video streaming framework of MPEG-4 standard. Based on FGS video coding framework which had relatively low coding efficiency, this paper proposed a novel video coding framework on the basis of leaky prediction techniques. Introduced leaky prediction factors into the improved FGS coding framework. Adaptively formed the temporal prediction by using enhancement layer reference and the base layer reference utilizing coded information in the base layer. The proposed theme could remove cumulative error and improve coding efficiency. The experimental results show that coding efficiency is improved by about 4dB than FGS under the same expermental conditions.