Analysis Operator Learning and its Application to Image Reconstruction

DOI10.1109/TIP.2013.2246175MaRDI QIDQ5373393zbMATH OpenOpenAlexDBLPWikidataFDO

Authors Simon Hawe, Klaus J. Diepold, Martin Kleinsteuber

Publication date 27 October 2017

Published in IEEE Transactions on Image Processing (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1204.5309

Mathematics Subject Classification ID

Directional data; spatial statistics (62H11) Estimation in multivariate analysis (62H12) Computing methodologies for image processing (68U10) Image processing (compression, reconstruction, etc.) in information and communication theory (94A08) Signal theory (characterization, reconstruction, filtering, etc.) (94A12)

Abstract: Exploiting a priori known structural information lies at the core of many image reconstruction methods that can be stated as inverse problems. The synthesis model, which assumes that images can be decomposed into a linear combination of very few atoms of some dictionary, is now a well established tool for the design of image reconstruction algorithms. An interesting alternative is the analysis model, where the signal is multiplied by an analysis operator and the outcome is assumed to be the sparse. This approach has only recently gained increasing interest. The quality of reconstruction methods based on an analysis model severely depends on the right choice of the suitable operator. In this work, we present an algorithm for learning an analysis operator from training images. Our method is based on an $el{l}_p$-norm minimization on the set of full rank matrices with normalized columns. We carefully introduce the employed conjugate gradient method on manifolds, and explain the underlying geometry of the constraints. Moreover, we compare our approach to state-of-the-art methods for image denoising, inpainting, and single image super-resolution. Our numerical results show competitive performance of our general approach in all presented applications compared to the specialized state-of-the-art techniques.

Cited in

(18)

• Metrically regular vector field and iterative processes for generalized equations in Hadamard manifolds
• Analysis sparse representation for nonnegative signals based on determinant measure by DC programming
• Structured overcomplete sparsifying transform learning with convergence guarantees and applications
• A bimodal co-sparse analysis model for image processing
• Hybrid Riemannian conjugate gradient methods with global convergence properties
• Riemannian stochastic fixed point optimization algorithm
• Manifold optimization-based analysis dictionary learning with an \(\ell_{1/2}\)-norm regularizer
• \(\ell^1\)-analysis minimization and generalized (co-)sparsity: when does recovery succeed?
• A new algorithm framework for image inpainting in transform domain
• Theory and application of connectivity maskbased reconstruction opening operator
• Sparsity and nullity: paradigms for analysis dictionary learning
• First order methods for optimization on Riemannian manifolds
• Cosparsity in Compressed Sensing
• Discriminative Nonlinear Analysis Operator Learning: When Cosparse Model Meets Image Classification
• An extragradient-type algorithm for variational inequality on Hadamard manifolds
• An introduction to continuous optimization for imaging
• Bilevel Methods for Image Reconstruction
• Sampling in the analysis transform domain

This page was built for publication: Analysis Operator Learning and its Application to Image Reconstruction

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5373393)