ON THE POCS-BASED POSTPROCESSING TECHNIQUE TO REDUCE THE BLOCKING ARTIFACTS IN TRANSFORM CODED IMAGES

In this paper, we propose a novel postprocessing technique, based on t he theory of projections onto convex sets (POCS), to reduce the blocki ng artifacts in transform-coded images. It is assumed, in our approach , that the original image is highly correlated. Thus, the global frequ ency characteristics in two adjacent blocks are similar to the local o nes in each block. We consider the high-frequency components in the gl obal characteristics of a decoded image, which are not found in the lo cal ones, as the results from the blocking artifact. We employ N-point discrete cosine transform (DCT) to obtain the local characteristics, and 2N-point DCT to obtain the global ones, and then derive the relati on between N-point and 2N-point DCT coefficients. A careful comparison of N-point with 2N-point DCT coefficients makes it possible to detect the undesired high-frequency components, mainly caused by the blockin g artifact. Then, we propose novel convex sets and their projection op erators in the DCT domain. The performances of the proposed and conven tional techniques are compared on the still images, decoded by JPEG, T he results show that, regardless of the content of the input images, t he proposed technique yields significantly better performance than the conventional techniques in terms of objective quality, subjective qua lity, and convergence behavior.