Neural‐network‐based regularization methods for inverse problems in imaging

A Habring, M Holler - GAMM‐Mitteilungen, 2024 - Wiley Online Library
This review provides an introduction to—and overview of—the current state of the art in
neural‐network based regularization methods for inverse problems in imaging. It aims to …

Learning from small data sets: Patch‐based regularizers in inverse problems for image reconstruction

M Piening, F Altekrüger, J Hertrich… - GAMM …, 2024 - Wiley Online Library
The solution of inverse problems is of fundamental interest in medical and astronomical
imaging, geophysics as well as engineering and life sciences. Recent advances were made …

Posterior sampling based on gradient flows of the MMD with negative distance kernel

P Hagemann, J Hertrich, F Altekrüger, R Beinert… - arXiv preprint arXiv …, 2023 - arxiv.org
We propose conditional flows of the maximum mean discrepancy (MMD) with the negative
distance kernel for posterior sampling and conditional generative modeling. This MMD …

[图书][B] Generalized normalizing flows via Markov chains

PL Hagemann, J Hertrich, G Steidl - 2023 - cambridge.org
Normalizing flows, diffusion normalizing flows and variational autoencoders are powerful
generative models. This Element provides a unified framework to handle these approaches …

Training Adaptive Reconstruction Networks for Blind Inverse Problems

A Gossard, P Weiss - SIAM Journal on Imaging Sciences, 2024 - SIAM
Neural networks allow solving many ill-posed inverse problems with unprecedented
performance. Physics informed approaches already progressively replace carefully hand …

Manifold learning by mixture models of vaes for inverse problems

GS Alberti, J Hertrich, M Santacesaria… - Journal of Machine …, 2024 - jmlr.org
Representing a manifold of very high-dimensional data with generative models has been
shown to be computationally efficient in practice. However, this requires that the data …

Conditional generative models are provably robust: Pointwise guarantees for bayesian inverse problems

F Altekrüger, P Hagemann, G Steidl - arXiv preprint arXiv:2303.15845, 2023 - arxiv.org
Conditional generative models became a very powerful tool to sample from Bayesian
inverse problem posteriors. It is well-known in classical Bayesian literature that posterior …

Learning regularization parameter-maps for variational image reconstruction using deep neural networks and algorithm unrolling

A Kofler, F Altekrüger, F Antarou Ba, C Kolbitsch… - SIAM Journal on Imaging …, 2023 - SIAM
We introduce a method for the fast estimation of data-adapted, spatially and temporally
dependent regularization parameter-maps for variational image reconstruction, focusing on …

Inverse problem regularization with hierarchical variational autoencoders

J Prost, A Houdard, A Almansa… - Proceedings of the …, 2023 - openaccess.thecvf.com
In this paper, we propose to regularize ill-posed inverse problems using a deep hierarchical
Variational AutoEncoder (HVAE) as an image prior. The proposed method synthesizes the …

NF-ULA: Normalizing Flow-Based Unadjusted Langevin Algorithm for Imaging Inverse Problems

Z Cai, J Tang, S Mukherjee, J Li, CB Schönlieb… - SIAM Journal on Imaging …, 2024 - SIAM
Bayesian methods for solving inverse problems are a powerful alternative to classical
methods since the Bayesian approach offers the ability to quantify the uncertainty in the …