基于自适应量化矩阵的硬判决量化算法  

Novel hard decision quantization algorithm based on adaptive quantization matrix

作　　者：李瑞阳[1] 王鸿奎[1] LI Ruiyang;WANG Hongkui(China Jiliang University, Hangzhou 310018, China)

机构地区：[1]中国计量大学,浙江杭州310018

出　　处：《电视技术》2017年第11期12-18,共7页Video Engineering

摘　　要：在视频编码中,视频量化一般分为硬判决量化(HDQ)和软判决量化(SDQ),HDQ与SDQ相比,编码性能虽有所损失,但其编码复杂度低,易于硬件实现的优点依旧是主流编码器所主要采用的量化算法。人眼具有对图像中的高频细节不敏感的特性。因此,基于Bayes最小误判概率约束,离线构建基于视频内容自适应的量化矩阵,在模拟感知SDQ算法机理下,对高频低频分量采用不同的量化步长,提高视频的主观质量和HDQ算法性能。仿真实验表明,相比于传统的HDQ算法,该文算法能达到平均5.048%的码率节省,其中WVGA和WQVGA格式平均达到10.65%的码率节省。相比于感知SDQ算法,平均码率增加仅有1.464%;算法复杂度方面,编码一帧的时间相比于感知SDQ节省了32.956%。In video coding,video quantization is generally divided into hard decision quantization （HDQ） and soft decision quantization （SDQ）,the coding performance of HDQ is lost compared to SDQ,but it still is the mainstream encoder used by the quantitative algorithm with the advantages that a low coding complexity and the application on the hardware,The human eye has a characteristic that is insensitive to high frequency detail in the image.Therefore,according to the Bayesian minimum error probability constraint,the quantization matrix based on video content adaptive is constructed.Under the simulated soft decision quantization mechanism,different quantization steps are adopted for high frequency and low frequency components,as a result,the subjective quality and the performance of HDQ algorithm are strengthened.Simulation results show that compared with the traditional HDQ algorithm,the algorithm can achieve an average saving of 5.048％ bit rate,of which WVGA and WQVGA format gain an average saving to 10.65％ bit rate.Compared with the perceived SDQ algorithm,the average bit rate increases by only 1.464％.In terms of algorithm complexity,the time to encode a frame is 32.956％ savings compared to the perceived SDQ.

关 键 词：视频编码 软判决量化 硬判决量化 量化矩阵 

分 类 号：TN919.81[电子电信—通信与信息系统]