面向HEVC的Golomb-Rice编码参数自适应策略

HEVC-oriented Golomb-Rice Parameter Adaptive Strategy

为了进一步降低高性能视频编码(HEVC)的输出码率,针对基于上下文的自适应算术编码(CABAC),提出了一种改进算法。利用大尺寸变换单元(TU)和变换跳过模式系数块大值系数较多的特点,首先在32×32变换单元中,根据已编码4×4系数组(CG)系数值的分布特性,自适应决定下一个CG的哥伦布-莱斯(Golomb-Rice)初始参数值;其次,在变换跳过模式系数块中设置初始Golomb-Rice参数为1,再利用相邻系数的相关性,根据已编码系数绝对值大小自适应决定下一系数的编码参数值。实验结果表明,与HEVC标准算法HM16.0相比,所提算法能达到0.09%~2.75%的比特率下降,平均有效率90%以上,且峰值信噪比(PSNR)无损失,编码时间平均只增加了0.08%。与代表文献相比,所提算法平均节省0.49%比特率,PSNR平均提高0.01 d B。

In order to further reduce high efficiency video coding（ HEVC） output rate,this paper presents an improved algorithm of Context-based Adaptive Arithmetic Coding（ CABAC）. Firstly,by using the feature that the large size transform unit（ TU） and the transform skip mode coefficient block has more large coefficients in CABAC,according to the distribution characteristics of the coded 4 ×4 coefficients array（ CG） in32 ×32 transform unit,the initial Golomb-Rice parameter of the next CG is adaptively determined. Then,in the transform skip mode coefficient block,the initial Golomb-Rice parameter is set to 1,and then the coding parameter of the next coefficient is adaptively determined according to the absolute value of the coded coefficients. The experimental results show that,compared with the HEVC standard algorithm HM16. 0,the proposed algorithm brings down the bit rate by 0. 09% - 2.75%,the average efficiency is more than 90%,peak signal-to-noise ratio（PSNR） has no loss and coding time is only increased by 0.08%. Compared with the representative literature,the proposed algorithm can save the average bit rate by 0.49%,and PSNR is increased by an average of 0.01 d B.