Suboptimality of Nonlocal Means for Images with Sharp Edges
Lists
Only you see this
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 log 1/2+ n, for an n-pixel image with > 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 .
BibTeX
@inproceedings{maleki2012suboptimality,
author = {Arian Maleki and Manjari Narayan and Richard G. Baraniuk},
title = {{Suboptimality of Nonlocal Means for Images with Sharp Edges}},
booktitle = {Applied and Computational Harmonic Analysis},
year = {2012},
doi = {10.1016/j.acha.2012.02.003},
}| Anisotropic Nonlocal Means Denoising | arXiv · 2011 |
| Oracle Inequalities and Minimax Rates for Nonlocal Means and Related Adaptive Kernel-Based Methods | SIAM J. Imaging Sci · 2012 |
| Non-Local Means Denoising | Image Process. Line · 2011 |
| Blockwise SURE Shrinkage for Non-Local Means | arXiv · 2013 |
| On Nearest Neighbors in Non Local Means Denoising | arXiv · 2017 |
| Wedgelets: nearly minimax estimation of edges | The Annals of Statistics · 1999 |