Most of the transform-based image compression techniques produce visible artifacts in the reconstructed image, particularly at low bit rates. In this paper, we propose an iterative technique to reduce the unwanted degradations such as blocking artifacts while keeping the necessary detail present in the original image. The proposed technique makes use of a priori information about the original image by using a nonstationary Gauss-Markov model. A MAP estimate is obtained iteratively using mean field annealing. A additional a priori information about the transform of the original image is incorporated into the estimation process by projecting the image onto a set defined by the quantizer at each iteration. The performance of the proposed algorithm was tested on JPEG compressed images. It is shown to be effective in removing the coding artifacts present in the low bit rate compressed images.
|Original language||English (US)|
|Journal||ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings|
|State||Published - Jan 1 1994|
ASJC Scopus subject areas
- Signal Processing
- Electrical and Electronic Engineering