摘要
提出了一种衡量帧内模式方向预测能力的新准则,指出H.264/AVC标准中帧内预测模式在不同方向上的预测能力存在非均衡性,因而限制了某些预测模式下帧内预测性能的发挥.为克服这一缺陷,提出了一种H.264的帧内预测扩展模式和一种灵活更改4×4编解码次序的策略.理论分析和实验结果表明,所提出的方法在没有增加新语法元素的同时,使得帧内预测的方向均衡性得到极大的改善,最弱模式的预测能力提高了30%以上,并获得了较为显著的编码增益,甚至超过了使用H.264的均值预测模式所带来的增益.
Unbalance among the intra mode prediction abilities in H. 264/AVC standard was unveiled by utilizing a criterion to measure the prediction ability of directional intra prediction modes. In order to mitigate the unbalance, which weakens the intra prediction, an extension of H. 264 intra prediction modes and a flexible encode/decode reordering strategy were proposed, and a more isotropic intra prediction free of any additional syntax element was realized by applying them. The weakest intra mode is enhanced by 30%, and a more remarkable coding gain is achieved by using the proposed methods, which is greater than that by using intra DC mode in H. 264.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2006年第5期717-723,共7页
Journal of Shanghai Jiaotong University
基金
上海市自然科学基金资助项目(04ZR14082)
