Statistical multiresolution Dantzig estimation in imaging: Fundamental concepts and algorithmic framework
Authors:
- Klaus FrickKFKlaus Frick
- Philipp MarnitzPMPhilipp Marnitz
Open Access
- 1 January 2012
- journal article
- Published by Institute of Mathematical Statistics in Electronic Journal of Statistics
- Vol. 6 (none) , 231-268
Abstract
In this paper we are concerned with fully automatic and locally adaptive estimation of functions in a “signal + noise”-model where the regression function may additionally be blurred by a linear operator, e.g. by a convolution. To this end, we introduce a general class of statistical multiresolution estimators and develop an algorithmic framework for computing those. By this we mean estimators that are defined as solutions of convex optimization problems with ℓ∞-type constraints. We employ a combination of the alternating direction method of multipliers with Dykstra’s algorithm for computing orthogonal projections onto intersections of convex sets and prove numerical convergence. The capability of the proposed method is illustrated by various examples from imaging and signal detection.Keywords
All Related Versions
This publication has 43 references indexed in Scilit:
- Regularization of ill-posed linear equations by the non- stationary augmented Lagrangian methodJournal of Integral Equations and Applications, 2010
- Shao's theorem on the maximum of standardized random walk increments for multidimensional arraysESAIM: Probability and Statistics, 2009
- Consistencies and rates of convergence of jump-penalized least squares estimatorsThe Annals of Statistics, 2009
- Discussion: The Dantzig selector: Statistical estimation when p is much larger than nThe Annals of Statistics, 2007
- Confidence Bands for Isotonic Median Curves Using Sign TestsJournal of Computational and Graphical Statistics, 2004
- Multiscale Testing of Qualitative HypothesesThe Annals of Statistics, 2001
- Adaptive Weights Smoothing with Applications to Image RestorationJournal of the Royal Statistical Society Series B: Statistical Methodology, 2000
- Convergence of an Iterative Method for Total Variation DenoisingSIAM Journal on Numerical Analysis, 1997
- Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopyOptics Letters, 1994
- Why Least Squares and Maximum Entropy? An Axiomatic Approach to Inference for Linear Inverse ProblemsThe Annals of Statistics, 1991