K.Sivakami Sundari, V. Sadasivam

Professor and Head, Department of Information Technology
PSNA College of Engineering and Technology, Dindigul, INDIA,
Professor and Head, CSE Department, MS University, Tirunelveli

Abstract:

Many multimedia applications require image compression with high compression ratio to overcome the difficulties in dealing huge volume of image data. At high compression ratios, the error introduced by quantization of the transform coefficients produces visually undesirable patterns known as compression artifacts that dramatically lower the perceived quality of a particular image. A great deal of effort has been invested in attempts to solve this problem while preserving the information content of the image. Proposed work primarily concentrates on the blocking artifacts of JPEG images and to a degree over the ringing artifacts of JPEG 2000 images. There exist three different approaches to reduce the artifacts as Preprocessing, Post processing and Transform domain techniques. Recently, attention is diverted to optimize the solution. Current work computes the measure of blocking artifacts with the new parameter named as Total blocking Error. Efficient suppression of artifacts can be controlled by the scaling parameter in the Quantisation process, and by the kernel in the filtering process. Genetic Algorithm (GA) is one of the emerging optimization techniques. Hence an attempt is made to optimize the kernel of the filter and the scaling parameter of the quantization with GA. A spatial domain algorithm can enhance further the quality of the image by preserving fine details. Dynamic Range Processing divide the image into luminance and chrominance component and converted to a reduced range with logarithmic mapping. Solving a Poisson equation on the attenuated modified gradient field preserves fine details. Finally the integrated in formations are remapped to the original dynamic range with inverse logarithm.

Keywords: Genetic Algorithm, Artifact reduction, DCT, Adapted Q etc.

BibTex:

@ARTICLE{P1150649006,

AUTHOR = {K.Sivakami Sundari and V. Sadasivam},

TITLE = {Reduction of Artifacts in JPEG images with Genetic Algorithm and Boundary pixel replacement},

JOURNAL ={ICGST International Journal on Graphics, Vision and Image Processing},

YEAR = {2007},

VOLUME = {07},

ISSUE ={1},

PAGES={17--24}

}

(Full Paper 767 KB)