The robustness of proximal penalty algorithms in restoration of noisy image
Abstract
The nondifferentiable convex optimization has an importance crucial in the image restoration for this and in this article we present the performance of the Prox method adapted to the restoration of noisy images. Following of our article ([12]), we illustrate in this work thesuperior efficacy of this algorithm “Prox” ([12]) then we are comparing the obtained numerical results with the algorithms of Wiener filtering ([7], [16]), total variation ([5]) and wavelet soft-thresholding denoising ([1], [12], [13]), in terms of image quality and convergence.
Our first experiments showed that by applying of Prox algorithm for restoration of noised image by the white Gaussian noise we obtain a top results of denosed image with high quality (net, not rehearsed and unsmoothed; textures are preserved) in addition to the convergence of the algorithm is ensured whatever the values of SNR.
Keywords
References
- S. Anthoine et Al, Some proximal methods for Poisson intensity CBCT and PET, Inverse Problems and Imaging, Vol. 6, Number 4 (2012).
- A. Auslender et Al, Penalty-Proximal in convex programming, J.O.T.A., Vol. 55 (1987), pp. 1-21.
- G. Aubert and P. Kornprobst, Mathematical Problems in Image Processing, volume 147 of Applied Mathematical Sciences, Springer-Verlag, 2002.
- F. Bach et Al, Optimization with sparsity-inducing, Foundations and Trends in Machine Learning, Vol. 4, Number 1 (2012), pp. 1-106.
- M. Bergounioux, Quelques méthodes mathématiques pour le traitement d’image, Notes de cours, 2009.
- A. Chambolle, An algorithm for total variation minimization and applications, Jour. Math. Imag. Vision, 20 (2004), pp. 89-97.
- S.G. Chang, Bin Yu and M. Vetterli, Adaptive wavelet thresholding for image denoising and compression, IEEE Transactions on Image Processing, Vol. 9, Issue 9 (2002), pp. 1532-1546.
- D. Cremers et Al, Convex relaxation techniques for segmentation, stereo and multiview reconstruction, dans Markov Random Fields for Vision and Image Processing, MIT Press, 2011.
Details
Primary Language
English
Subjects
Mathematical Sciences
Journal Section
Research Article
Publication Date
December 1, 2017
Submission Date
November 23, 2016
Acceptance Date
January 14, 2017
Published in Issue
Year 2017 Volume: 46 Number: 6