Post-processing for removing coding artifacts using edge-preserving regularization

In low bit rate image coding, the perceived quality of the decoded image is severely degraded by various artifacts, such as blocking, blotchiness and ringing effects. We propose an effective post-processing algorithm with low computational complexity for removing coding artifacts. We first give an image recovery model for post-processing. Based on this model, a smooth constraint is imposed, together with an edge-preserving regularization. Experimental results have shown that the proposed algorithm can significantly suppress the coding artifacts and improve the perceived image quality.