基于上下文的自适应二进制算术编码研究

Research on context-based adaptive binary arithmetic coding

针对视频编码标准H.264中基于上下文的自适应二进制算术编码(CABAC)的复杂度高和存储空间大的问题,提出了一种简化的编码器模型.在对CABAC熵编码器进行理论分析后,给出了CABAC的编码效率公式,进而推导出不同概率时的理论最优量化点集合;通过调整CABAC中消除乘法运算的编码器的区间精度和倍乘系数,研究了这些参数和量化点集合对CABAC的编码效率和存储空间的影响,获得了该种简化模型,并利用此编码模型对不同特性的图像序列进行测试.仿真结果表明,该模型能够在编码效率降低不多的情况下,大大降低复杂度和存储空间,适用于对存储空间要求较高的视频通信.

To solve the problem of high computational complexity and large memory space in context-based adaptive binary arithmetic coding (CABAC) of the video coding standard H.264, a new simple encoder scheme was proposed. Theoretical analysis of CABAC entropy coding architecture yielded a formula for calculating the coding efficiency, and optimal quantization point sets for different probabilities were obtained from this formula. After adjusting the precision of the range and scaling factor in the multiplication-free adaptive binary arithmetic coder in CABAC, the effects of these parameters and the quantization point sets on the coding efficiency and memory space for the CABAC entropy coding were investigated, and the new encoder scheme was then set up. The experimental results on different picture sequences under this scheme show that the memory space and computational complexity can be reduced greatly without much decrease of the coding efficiency. This scheme is efficient in the video communication where less memory space is provided.