Yongjin Cho, C.-C. Jay Kuo, Ming Hsieh, Do-Kyoung Kwon
{"title":"GOP-based Rate Control for H.264/SVC with Hierarchical B-pictures","authors":"Yongjin Cho, C.-C. Jay Kuo, Ming Hsieh, Do-Kyoung Kwon","doi":"10.1109/IIH-MSP.2007.422","DOIUrl":null,"url":null,"abstract":"A GOP-based rate control algorithm for hierarchical B-pictures in H.264/SVC is proposed in this work. The problem is firstly formulated as a bit allocation problem among different temporal layers. Due to the dependency of these layers, we propose a Q-selection tree solution framework. The monotonicity property is exploited to prune the Q-selection tree. Then, the rate-distortion (R-D) data is calculated at a selected set of leaf nodes in any admissible Q-selection tree. The R-D data of other leaf nodes are interpolated using the log-linearity property of a local R-D model. It is shown by experiments that the proposed algorithm achieves a significant coding gain improvement at different target bit rates.","PeriodicalId":385132,"journal":{"name":"Third International Conference on Intelligent Information Hiding and Multimedia Signal Processing (IIH-MSP 2007)","volume":"301 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2007-11-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"8","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Third International Conference on Intelligent Information Hiding and Multimedia Signal Processing (IIH-MSP 2007)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/IIH-MSP.2007.422","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 8
Abstract
A GOP-based rate control algorithm for hierarchical B-pictures in H.264/SVC is proposed in this work. The problem is firstly formulated as a bit allocation problem among different temporal layers. Due to the dependency of these layers, we propose a Q-selection tree solution framework. The monotonicity property is exploited to prune the Q-selection tree. Then, the rate-distortion (R-D) data is calculated at a selected set of leaf nodes in any admissible Q-selection tree. The R-D data of other leaf nodes are interpolated using the log-linearity property of a local R-D model. It is shown by experiments that the proposed algorithm achieves a significant coding gain improvement at different target bit rates.