Variational networks: an optimal control approach to early stopping variational methods for image restoration

DOI10.1007/s10851-019-00926-8zbMath1434.68626arXiv1907.08488OpenAlexW3011127553WikidataQ92000462 ScholiaQ92000462MaRDI QIDQ1988355

Thomas Pock, Erich Kobler, Alexander Effland, Karl Kunisch

Publication date: 23 April 2020

Published in: Journal of Mathematical Imaging and Vision (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1907.08488

zbMATH Keywords

gradient flowvariational problemsoptimal control theorydeep learningearly stoppingvariational networks

Mathematics Subject Classification ID

Artificial neural networks and deep learning (68T07) Computing methodologies for image processing (68U10) Image processing (compression, reconstruction, etc.) in information and communication theory (94A08) Control problems involving ordinary differential equations (34H05) Numerical methods for initial value problems involving ordinary differential equations (65L05) Optimality conditions for problems involving ordinary differential equations (49K15)

Related Items

WARPd: A Linearly Convergent First-Order Primal-Dual Algorithm for Inverse Problems with Approximate Sharpness Conditions ⋮ Deep Learning--Based Dictionary Learning and Tomographic Image Reconstruction ⋮ Convergence of a Piggyback-Style Method for the Differentiation of Solutions of Standard Saddle-Point Problems ⋮ Bayesian Imaging Using Plug & Play Priors: When Langevin Meets Tweedie ⋮ WPPNets and WPPFlows: The Power of Wasserstein Patch Priors for Superresolution ⋮ Explainable bilevel optimization: an application to the Helsinki Deblur Challenge ⋮ Deep learning via dynamical systems: an approximation perspective ⋮ Convolutional proximal neural networks and plug-and-play algorithms ⋮ RICAM, the Johann Radon Institute for Computational and Applied Mathematics ⋮ Nonlinear Power Method for Computing Eigenvectors of Proximal Operators and Neural Networks ⋮ Learnable Descent Algorithm for Nonsmooth Nonconvex Image Reconstruction ⋮ Shared Prior Learning of Energy-Based Models for Image Reconstruction ⋮ Bilevel Methods for Image Reconstruction

Uses Software

• Regularization tools
• BSDS
• IR Tools

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Nonlinear total variation based noise removal algorithms
• A Fast Iterative Shrinkage-Thresholding Algorithm for Linear Inverse Problems
• Fields of experts
• Iterative regularization methods for nonlinear ill-posed problems
• Nonlinear eigenproblems in image processing and computer vision
• Deep learning as optimal control problems: models and numerical methods
• A mean-field optimal control formulation of deep learning
• A proposal on machine learning via dynamical systems
• IR tools: a MATLAB package of iterative regularization methods and large-scale test problems
• Boosting with early stopping: convergence and consistency
• On early stopping in gradient descent learning
• On the Landweber iteration for nonlinear ill-posed problems
• Inertial Proximal Alternating Linearized Minimization (iPALM) for Nonconvex and Nonsmooth Problems
• Lagrange Multiplier Approach to Variational Problems and Applications
• Insights Into Analysis Operator Learning: From Patch-Based Sparse Models to Higher Order MRFs
• Stable architectures for deep neural networks
• On the second-order asymptotical regularization of linear ill-posed inverse problems
• Discrete Inverse Problems
• An introduction to continuous optimization for imaging
• An Iteration Formula for Fredholm Integral Equations of the First Kind
• Numerical Methods for Ordinary Differential Equations