Blocking artifact reduction in compressed images based on edge-adaptive quadrangle meshes

摘要

During image deblocking, the preservation of visually significant edges in the image is important. Based on the observation that blocking artifact along edge direction is difficult for human perception, a number of existing techniques have proposed to filter the edge pixels with low weight or bypass them altogether to avoid blurring. However, the un-processed artifacts that are anti-parallel to the edge direction often fragment the edges and seriously degrade visual quality. In this paper, by considering the behavior of intensity evolution along and across edges, we propose a new POCS-based algorithm using a new smoothness constraint set. The new constraint set is realized indirectly by limiting the distance between the coded image and the image simulated using edge-adaptive quadrangle meshes. Experimental results indicated that our method compares favorably with several existing techniques.