H.264/AVC零值量化DCT系数分级预判算法

Classified detection algorithm of zero-quantized DCT coeffcients for H.264/AVC

在H.264/AVC标准算法中,运动补偿后的残差块经过4×4整数离散余弦变换(DCT)及量化过程后,往往出现大量的零值量化系数(ZQDCT),而这些零值系数完全可以通过提前判别来省略其变换量化过程.为了减少ZQDCT的冗余计算,在分析提取出最佳全零块判决阈值的基础上,结合高斯分布模型和DCT系数能量"Z"形分布特性,提出了基于该新阈值的零值量化系数分级预判算法,在残差块DCT变换之前按分级模式预判出ZQDCT,从而跳过其变换量化操作.实验结果表明,与标准算法相比,该算法明显降低了编码数据量,同时减少了最高达42%的离散余弦变换(DCT)、量化(Q)、反量化(IQ)以及离散余弦反变换(IDCT)过程计算时间.

In H.264/AVC video coding, there are always a large number of zero-quantized DCT （ZQDCT） coefficients obtained from 4×4 integer discrete cosine transform （DCT） and quantization process with the residual blocks of motion compensation, which can be detected in advance and set zero before the transformation and quantization process. Based on the extraction of the best all-zero block detection threshold, combined with Gaussian model and ＂Z＂ shape characteristic of DCT energy distribution, a classified detection algorithm of zero-quantized DCT coeffcients is proposed to reduce the redundant calculation of ZQDCT. In the algorithm, the residual block coeffcients can be detected as ZQDCT before DCT transformation according to the classification model, which can skip the transform quantization operation. Experimental results show that, compared with the standard algorithm, the proposed algorithm reduces the encoding data bit rate obviously, and achieve a highest timesaving of 42 % during the process of the discrete cosine transform （DCT）, quantization （Q）, inverse quantization （IQ） and inverse discrete cosine transform （IDCT）.