基于帧内预测模式的HEVC音视频同步算法  

Audio-video Synchronization Algorithm for HEVC Based on Intra Prediction Mode

作　　者：徐辉[1] 王晓东[1] 王让定[1] 章联军[1] 

机构地区：[1]宁波大学信息科学与工程学院,浙江宁波315000

出　　处：《计算机工程》2015年第12期241-248,共8页Computer Engineering

基　　金：国家自然科学基金资助重点项目(U1301257);国家科技支撑计划基金资助项目(2012BAH67F01);浙江省教育厅科研计划基金资助项目(Y201327703);浙江省科技厅创新团队自主设计基金资助项目(2012R10009-08);宁波市科技创新团队研究计划基金资助项目(2011B81002)

摘　　要：针对目前H.264嵌入式音视频同步算法在高清、超高清视频应用方面的局限性,提出一种基于帧内预测模式的高效视频编码音视频同步方法。在帧内预测过程中,选取纹理复杂度较高的4×4亮度块调制预测模式,在分析相邻预测模式相关性的基础上,对35种预测模式进行分组,通过改进可变长编码算法及其映射规则,将音频信息进行动态分组,建立预测模式组与变长码组之间的动态双映射关系。为提高视频质量,对修改后的预测模式进行重编码。在解码端,只需将解码得到的预测模式根据双映射关系进行音频信息的提取及重构。仿真结果表明,该算法可保证音频数据的正确性和完整性,实现高清、超高清视频的音视频同步编码传输。Due to the limitations of embedded audio-video synchronization algorithm based on H.264,when applying in HD,U-HD video,an audio-video synchronization algorithm is proposed for High Efficiency Video Coding（HEVC）based on intra prediction.This paper modifies prediction modes by selecting 4×4 luminance blocks with high-complexity texture during the process of prediction coding.The 35 prediction modes are divided into groups on the basis of analyzing the correlation of prediction modes.And by improving the variable length coding algorithms and mapping rules,the audio data is divided into groups dynamically,thus the dynamic dual mapping between the prediction modes and variable length code groups is established.For the purpose of reducing the impact of modifying the prediction quality of video,recoding the prediction mode revised,at the decoder,audio extraction and reconstruction just depends on dual mappings between the decoded prediction modes and audio code groups.Experimental results show that the proposed scheme can ensure the validity and integrity of the audio data and it can achieve synchronization effectively and quality degradation in HD,U-HD video.

关 键 词：帧内预测 高效视频编码 相关性 变长编码 音视频同步 

分 类 号：TN919.8[电子电信—通信与信息系统]