视频编码预处理算法研究

Pre-processing algorithm for video coding

针对现有降噪采用时空域滤波、去块效应采用环路滤波/后处理这种"分而治之"方案的缺点,通过分析降噪的基本技术、低码率视频应用中块效应产生的原因和去块效应的常用方法,从滤除不需要或相对不重要的高频离散余弦变换(DCT,discrete cosine transform)系数角度提出了降噪和去块效应可同时处理的设想。新提出的预处理算法是根据图像纹理特性、运动情况和码率约束等,在图像的不同区域采用不同强度的自适应双边滤波,这样不仅去噪,滤除不重要的细节以便于高效压缩,避免块效应的产生,同时也可起到一定程度的码率控制效果。

After analyzing the disadvantage of separate solutions which uses spatial filtering to denoise and in-loop filtering/post-processing to deblock,we decribe a notion that denoising and deblocking are alike in terms of removing the high frequency information which are not required or less important.Moreover,in order to reduce coding complexity a pre-processing algorithm is proposed which can be used to filter noise and less important texture simultaneously.This pre-processing algorithm can adapt its filter strength according to the characteristic of both texture and object motion.Experimental results show that the new proposed algorithm improves the compression efficiency and avoids blocking artifacts,and it is also helpful to control bit rate to some extent.