论文部分内容阅读
网络的带宽抖动和数据传输错误是影响视频传输的重要原因。该文提出了一种新的多参考帧图像预测的视频编码算法,即多参考预测链,使视频具有时域可扩展性和抗错性解决网络的这两个潜在问题。该算法中P帧之间的参考预测关系构成多条链状结构,使视频具有时域可扩展性。链间的不相关性使每个链可独立解码,从而增强了码流抗错性。该算法采用R-D框架动态的管理预测关系链,从而适合网络带宽的变化而获得最佳的终端视觉效果。仿真实验表明,采用多参考预测链编码的码流时域可扩展性的应用非常可行,且表现出较高的抗错性。
Network bandwidth jitter and data transmission errors affect the video transmission is an important reason. In this paper, a new video coding algorithm for multi-reference frame prediction is proposed, that is, a multi-reference prediction chain. These two potential problems make the video have time-domain scalability and anti-error. The reference predictive relationship between P frames in this algorithm constitutes a multi-chain structure, which makes the video have time-domain extensibility. The irrelevance between the chains enables each chain to be independently decoded, thereby enhancing the error-correcting performance of the code stream. The algorithm uses the R-D framework to dynamically manage the prediction relationship chain, so as to adapt to the change of network bandwidth and obtain the best terminal visual effect. Simulation results show that the application of multi-reference prediction chain coding in the time-domain scalability is feasible and shows a good anti-error performance.