基于图像纹理特征的H.264/AVC帧内预测快速算法
Fast Intra Prediction Based on the Texture Feature for H.264/AVC Encoding
-
摘要: 为了降低H.264视频编码中帧内预测算法的复杂度,在基于4×4子块图像纹理检测的基础上,提出了结合块类型选择和候选模式选择的两步帧内预测快速算法.利用边缘检测模板得到宏块中每个4×4子块的纹理特征后,首先基于各个子块的边缘强度之间的离差选择合适的预测块类型;再针对所选块类型,基于块内所属的4×4子块的边缘方向进一步选择候选模式.实验结果表明,与原始全搜索算法相比,该快速算法能在保证编码性能基本不变的前提下,对全I帧序列的编码时间可以缩短平均74%.Abstract: Intra coding in H.264/AVC can significantly improve the compression efficiency but at the cost of high computational complexity.To reduce the complexity,this paper presents a fast intra prediction algorithm including block type decision and candidate modes decision based on the texture detection of 4x4 blocks.After the texture feature of every 4x4 subblock in a macroblock by template matching is achieved,the proper block type is firstly chosen based on the deviation of subblocks'edge strengths among a macrock Then for the chosen block type,the candidate modes are further decided based on the edge direction of current blocks.Experimental results show that the proposed fast algorithm can achieve 74% time saving on average for encoding the all intra-frame sequence with little loss of coding performance with the reference software.
下载:
