
NuHAG :: TALKS
Talks given at NuHAG events






Feature Preserving Image Denoising using Interandintrascale Dependencies of Wavelet Coefficients Serena Papi Department of mathematics, University of Bologna ITALY given at strobl07 (17.06.07) id: 728 length: min status: accepted type: poster LINKPresentation: ABSTRACT:
\documentclass[11pt,a4paper]{article}
\usepackage{graphicx,float,amssymb,amsmath,array}
\renewcommand{\baselinestretch}{1}
\usepackage{a4wide}
\begin{document}
%\title{Edge Preserving Denoising via Anisotropic Wavelet Thresholding}
\title{Feature Preserving Image Denoising using Interandintrascale Dependencies of Wavelet Coefficients}
\author{D.Lazzaro,\; L.B.Montefusco, \; S.Papi}
\maketitle{}
\par\noindent
A new wavelet shrinkage algorithm for edge preserving image denoising is proposed. It represents an improvement of a previously
presented thresholding method \cite{MONTE} that adds to the denoising process the construction of the map of image edges.
The map is obtained by taking into consideration both spacescale dependencies and interscale correlations between corresponding coefficients
of the dyadic Mallat's wavelet decomposition of the image.
The knowledge of the location and the orientation of the edges allows us to perform an anisotropic shrinkage by killing the gradient components
along the edges while keeping or only slightly shrinking the orthogonal components. The nonedge coefficients are then shrinked using a
correlated PeronaMalik rule.
\par\noindent
The resulting anisotropic wavelet thresholding scheme succeeds in producing a denoised version of the original image that still presents
welldefined edges, while the noise components are removed.
\begin{thebibliography}{99}
\bibitem{MONTE} D. Lazzaro, L.B.Montefusco, EdgePreserving Wavelet Thresholding for Image Denoising,
J. Computat. Appl. Mat. (2006), doi:10.1016/j.cam.2006.10.069
\end{thebibliography}
\end{document}
