Suboptimality of nonlocal means for images with sharp edges

DOI10.1016/J.ACHA.2012.02.003zbMATH Open1264.94023arXiv1111.5867OpenAlexW1988175255MaRDI QIDQ448488FDOQ448488

Manjari Narayan, Richard G. Baraniuk, Arian Maleki

Publication date: 6 September 2012

Published in: Applied and Computational Harmonic Analysis (Search for Journal in Brave)

Abstract: We conduct an asymptotic risk analysis of the nonlocal means image denoising algorithm for the Horizon class of images that are piecewise constant with a sharp edge discontinuity. We prove that the mean square risk of an optimally tuned nonlocal means algorithm decays according to $n^{-1}lo{g}^{1/2+epsilon}n$, for an $n$-pixel image with $epsilon>0$. This decay rate is an improvement over some of the predecessors of this algorithm, including the linear convolution filter, median filter, and the SUSAN filter, each of which provides a rate of only $n^{-2/3}$. It is also within a logarithmic factor from optimally tuned wavelet thresholding. However, it is still substantially lower than the the optimal minimax rate of $n^{-4/3}$.

Full work available at URL: https://arxiv.org/abs/1111.5867

Recommendations

• Image edge sharpening via Heaviside substitution and structure recovery
• Image sharpening via Sobolev gradient flows
• Analysis of the Recovery of Edges in Images and Signals by Minimizing Nonconvex Regularized Least-Squares
• A new nonlocal variational setting for image processing
• Nonlocal Operators with Applications to Image Processing
• Image Selective Smoothing and Edge Detection by Nonlinear Diffusion
• Deblurring and Denoising of Images by Nonlocal Functionals
• Weighted nonlocal total variation in image processing
• A variational approach for sharpening high dimensional images

zbMATH Keywords

denoisingminimax risklinear filternonlocal meanswavelet thresholdinghorizon classSUSAN filter

Mathematics Subject Classification ID

Image processing (compression, reconstruction, etc.) in information and communication theory (94A08)

Cites Work

• Nonlinear total variation based noise removal algorithms
• A Review of Image Denoising Algorithms, with a New One
• Digital picture processing. An introduction. Transl. from the Russian
• Minimax estimation via wavelet shrinkage
• Optimally Sparse Multidimensional Representation Using Shearlets
• Title not available (Why is that?)
• Minimax theory of image reconstruction
• Sparse directional image representations using the discrete shearlet transform
• Wedgelets: Nearly minimax estimation of edges
• Anisotropic nonlocal means denoising

Cited In (4)

• Optimization and learning with nonlocal calculus
• Convergence theorems for the non-local means filter
• Nonlocal Means and Optimal Weights for Noise Removal
• Anisotropic nonlocal means denoising

This page was built for publication: Suboptimality of nonlocal means for images with sharp edges

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q448488)