Ringing Artifacts Reduction using Adaptive Morphological filter for JPEG2000

Abstract

The JPEG2000 is the latest presently image compression standard presented by International Stan-dards Organization which enables compressed image at very low bit rates. At a very low bit rate, the recon-structed image contains visible ringing artifacts at the image edge. In order to reduce the ringing artifacts, we propose a new filter for filtering the ringing artifacts and enhance the processed image quality. The proposed algorithm is performed on the encoder and the decoder of JPEG2000. In the encoder, the compressed image is partitioned using the DFQP (DWT Feature Quad-tree Partition) and morphological operations are conducted to select the optimal structure element in order to obtain the best filtering effect. Therefore, the side informa-tion used for selecting the structure element is coded using entropy coding and transmitted to the decoder. Through decoding by the morphological filter with the selected structure element, a filtered image with excel-lent quality is obtained. Simulation results demonstrate that the filtered image is enhanced by 0.95dB of PSNR.