可扩展视频编码中运动补偿时域滤波的改进  

IMPROVED MOTION COMPENSATED TEMPORAL FILTERING IN SCALABLE VIDEO CODING

作　　者：李俊[1] 刘晓冬[1] 

机构地区：[1]清华大学深圳研究生院,广东深圳518055

出　　处：《计算机应用与软件》2007年第4期93-95,共3页Computer Applications and Software

摘　　要：在基于小波变换的可扩展视频编码方案中,研究运动补偿时域滤波的优化方法。将有效的多相位子带内运动估计纳入分层运动估计策略中,用提升算法实现子带内运动补偿时域滤波,融入多假设预测以优化运动补偿技术;为了提高低码率下视频编码的可扩展性,提出更新算子的优化设计。实验结果显示,在保证运动估计精度下,合理选择时间滤波器,结合改善的提升实现中的更新操作,提高了编码系统的编码效率和可扩展性能。Based on scalable wavelet video coding scheme, optimization of motion compensated temporal filtering is studied. Combining with efficient multiphase motion estimation in hierarchical matching strategy, lifting implementation of in band MCTF is applied, by exploiting multihypothesis technology to optimize motion compensation;an optimized update operator in lifting implementation is proposed to improve scalability under low bit rate. Experiments show motion estimation accuracy can be obtained in proposed system, the appropriate selection of temporal filter and optimization of update operation in lifting, bring about improvement of video coding efficiency and scalability.

关 键 词：可扩展视频编码 子带内运动补偿时域滤波 提升实现 更新算子 

分 类 号：TN95[电子电信—信号与信息处理]