Sparse Decomposition Method for Image Denoising

EasyChair Preprint 4755

Abstract

The sparse respresentation of data  has been made an evolution on the part of the signal processing community. After defining the different notions related to sparseness, image denoising, and the problem of eliminating noise from medical images, we will explain the integration and contribution of this in different fields of application.

The general goal of sparse representations of data  is to find the best approximation of a target signal using a linear combination of a few elementary signals from a fixed collection. For this fact, several methods of the decomposition of the signal have been used like PCA (Principal Component Analysis), ICA (independent components analysis ), Duet method, MP (Matching pursuit), Least Angle Regression ( LARS),…. and some noise reduction schemes are offered. Schemas include DCT_OMP, DCT_BOMP, Log Gabor_BOMP, DCT_OCMP and Wavelet_OMP

In this paper,  we will use and measure the performance of   image denoising method using the   sparse representation and decimposition of images.

Keyphrases: The LARS (least angle regression) algorithm, The LASSO algorithm, The Orthogonal Matching Pursuit (OMP) algorithm, sparse Matching pursuit (MP)

@booklet{EasyChair:4755,
  author    = {Hatim Koraichi and Otman Chakkor},
  title     = {Sparse Decomposition Method for Image Denoising},
  howpublished = {EasyChair Preprint 4755},
  year      = {EasyChair, 2020}}