A Novel Blocking Effect Reduction Method Based on Human Visual System Features

In this paper, we present a brief study about the performance of nine post-processing methods on blocked images, and two new approaches which propose, respectively, the reduction of the blocking artifact and a measure for the visual perception of the Human Visual System (HVS) regarding this distortion. The studies concerning the chosen methods consist of detailed analysis about their methodology and their computational implementation. A new visual perception metric is introduced, which is called Blocking Visual Effect Index (BVEI). It is based on loss of correlation criteria in inter-blocks regions, distortion of local average luminance, and takes into account the contrast perception sensibility of human eye. The proposed post-processing method considers the features of HVS when varying the smoothing parameter inside the blocking areas in accordance with the perception of blocking artifacts. Thus, they are smoothed without loss of details. Aiming at obtaining a consistence with subjective evaluation of postprocessed images, this work provides comparative tests for evaluating the performance of the chosen methods and the proposed one, Also it emphasizes the efficiency of the new quality metric. This efficiency is evidenced by the increase of quality of post-processing images, regarding the test images.