基于相邻CU和率失真代价的HEVC帧内预测快速算法
Study on Fast Intra Prediction in HEVC Based on Adjaced CU and Rate-distortion Cost
本文针对HEVC帧内预测算法计算复杂度高的问题,提出了一种基于相邻CU和率失真代价的帧内预测快速算法。在帧内预测前利用相邻已编码CU的深度对当前CU的深度进行预判断;在帧内预测时利用当前CU的率失真代价与预先定义的阈值进行对比判断,跳过一些不适合所属区域内容的编码单元尺寸类型的编码过程。实验结果表明,在低复杂度(LoCo)和高效率(HE)两种编码配置下,本文所提快速算法与HM9.0相比,平均的压缩编码时间分别减少了27.04%和21.72%;而平均编码码率分别只增加了约0.33%和0.16%,平均亮度峰值信噪比分别只减少了0.028dB和0.021dB。另外,通过RD曲线和编码重构图像的主观比较发现本文所提的快速算法的编码性能与HM9.0相比,二者的编码性能几乎不存在差异,由此证明本文所提快速算法的有效性。
In this paper, we propose a fast intra prediction algorithm of HEVC based on the adjacent coding unit and rate-distortion cost to the problem of high computational complexity of intra prediction algorithm in HEVC. Before intra prediction, we introduce the depth relevance of the adjacent coding unit to preliminary judging the depth of the current coding unit. When coding the current coding unit, we skip some coding unit sizes that do not fit this region after contrast between their rate-distortion costs and the threshold. The experimental results show that the average encoding time of the proposed fast algorithm decreases by 27.04% and 21.72% respectively in Low Complexity and High Efficiency test conditions. However the average coding bit rate increases only by about 0.33% and 0.16% and the average brightness peak signal-to-noise ratio reduces 0.028dB and 0.021dB respectively. In addition, it can prove the validity of the proposed fast algorithm that the proposed fast algorithm almost has the same RD curves and reconstructed images comparing with HM9.0.
赵岩、陈贺新、樊倩
计算技术、计算机技术通信电子技术应用
帧内预测快速算法HEVC率失真代价U深度
Intra predictfast algorithmHEVCrate distortion costthe depth of CU
赵岩,陈贺新,樊倩.基于相邻CU和率失真代价的HEVC帧内预测快速算法[EB/OL].(2014-05-20)[2025-08-23].http://www.paper.edu.cn/releasepaper/content/201405-324.点此复制
评论