Objective: With the rapid development of cloud computing and mobile Internet technologies, screen image coding has become a new research hotspot in the field of video compression. The intra block copy (IBC) algorithm is a core algorithm in screen content coding (SCC) and has become an important part of screen content coding in standards such as high efficiency video coding (HEVC). In order to further eliminate the redundancy of displacement vector (DV) parameter coding in the IBC algorithm, a DV parameter coding algorithm is proposed according to the inherent intra-block matching characteristics and correlations of DV parameters.
Methods: Firstly, a DV prediction coding scheme combining neighboring blocks and the nearest block is adopted for the DV to be coded to further improve the prediction coding efficiency of DV; then, for DV with low prediction coding efficiency, an efficient DV direct coding scheme based on region division and adjustment is proposed. Three coding configurations of 17 test data sets in the SCC standard test data set are evaluated in terms of coding efficiency and complexity.
Results: The experimental results show that for the moving text and graphic sequence categories in the SCC standard test data set, under the three lossy coding configurations of all intra, random access and low delay, compared with the IBC algorithm in HEVC SCC, the proposed DV algorithm has an average reduction rate of BD (Bjϕntegaard Delta)-rate of the Y component of 1.04%, 0.87% and 0.93% respectively without any impact on the encoding and decoding complexity, and the reduction rate of BD-rate of the Y component under the all intra configuration can reach 2.99%.
Conclusion: The method in this paper is superior to the DV coding algorithm in the IBC algorithm in HEVC - SCC and can effectively improve the coding efficiency.
目的随着云计算和移动互联网技术的飞速发展,屏幕图像编码已成为视频压缩领域新的研究热点。帧内块复制(intra block copy,IBC)算法是屏幕内容编码(screen content coding,SCC)中的核心算法,已经成为高效视频编码(high efficiency video coding,HEVC)等标准中屏幕内容编码的重要组成部分。为了进一步消除IBC算法中的位移矢量(displacement vector,DV)参数编码的冗余,根据DV参数具有的固有帧内块匹配特性和相关性,提出了一种DV参数编码算法。方法首先,对待编码DV采用邻近块与最近块相结合的DV预测编码方案,进一步提高DV的预测编码效率;然后,对预测编码效率不高的DV,提出了一种基于区域划分与调整的高效DV直接编码方案。对SCC标准测试数据集合中的17个测试数据集合的3种编码配置从编码效率和复杂度两方面进行了评测。结果实验结果表明,对于SCC标准测试数据集合中移动的文字和图形序列类别,在全帧内、随机接入和低延迟这3种有损编码配置下,提出的DV算法与HEVC SCC中的IBC算法相比,在编解码复杂度没有任何影响的前提下,Y分量的BD(Bjϕntegaard Delta)-rate平均降低率分别为1. 04% 、0. 87%和0. 93%,全帧内配置下Y分量的BD-rate降低率可达2. 99% 。结论本文方法优于HEVC-SCC中的IBC算法中的DV编码算法,能有效提升编码效率。