Paper Details: Downloads: 506
Serial Number: P1180820002
Title: Image Restoration Approach Based on Wavelet Coefficients Modeling
Authors: Besma Miladi and M. Jérôme and N. Ellouze
Abstract: This paper presents the restoration approach based on the modeling of the wavelet coefficients of the natural images. Two algorithms are developed: the first algorithm developed concerns the modeling of the natural images using the wavelet coefficients; the second algorithm uses the modeling results for the image restoration. Simulations are operated on some standard test images degraded by an additive Gaussian noise with standard deviation s. The wavelet coefficients are obtained using redundant and unredundant wavelet decompositions. It is shown that the proposed approach gives the best results as well in objective quality as in subjective quality, compared to Wiener method and outperforms it significantly. The comparisons are also done with the J.Portilla et al method which is the best method in the literature; it is shown that our results are competitive with this method. Furthermore, we explicitly establish the correlation between the distribution of wavelet coefficients for natural images and the distribution of the same images degraded by an additive Gaussian noise, which is the new contribution of this work.
Keywords: Image modeling, Restoration, Wavelet coefficients.
Journal/Conference: International Journal of Digital Signal Processing
Volume: 8
Issue: 1
Submission Date: 5/1/2008 12:00:00 AM
Review Date: 7/1/2008 12:00:00 AM
Publishing Date: 12/1/2008 12:00:00 AM
Article Downloads: 506
Download:

Facebook