在基于块编码的图像压缩技术中,低比特率下重建图像的块边界会出现块效应.本文提出了一种有效的基于DCT域的块效应消除算法.首先通过DCT系数块的分解和合并,直接从两个相邻的系数块获取相应的转移块.然后对图像块的活动性进行分类,对于低活动性邻接块所构成的转移块,采用改进的线性函数替代阶梯函数的方法消除块效应,同时更新邻接块.最后采用后滤波和量化约束技术对更新块和高活动性块进行后处理.实验结果表明,本算法不仅有效去除块效应,而且能够保护图像的纹理和边缘.
In image compression techniques based on block coding,blocking artifacts occur near the block boundaries of the decoded images at low bit rates.In this paper,an efficient deblocking algorithm in DCT(discrete cosine transform) domain is proposed.The shifted block is obtained directly from its adjacent blocks via the decomposition and combination of DCT coefficient blocks.After block activity classification,for the shifted block between lowactivity adjacent blocks,the improved linear function is used to replace the step function to eliminate blocking artifacts,and then the shifted block is used to update its adjacent blocks.Finally,the post-filtering and quantization constraint techniques are adopted to post-process the updated and high-activity blocks.Experimental results show that the proposed approach can effectively reduce blocking artifacts and preserve the image textures and edges.