mlsanityData to 7 Oct 2026

Training Adaptive Reconstruction Networks for Blind Inverse Problems

arXiv 2202.11342 ↗First posted 23 Feb 2022cs.LG eess.IV eess.SPcited by 3

Not on the map. Why?

Abstract (arXiv)

Neural networks allow solving many ill-posed inverse problems with unprecedented performance. Physics informed approaches already progressively replace carefully hand-crafted reconstruction algorithms in real applications. However, these networks suffer from a major defect: when trained on a given forward operator, they do not generalize well to a different one. The aim of this paper is twofold. First, we show through various applications that training the network with a family of forward operators allows solving the adaptivity problem without compromising the reconstruction quality significantly.Second, we illustrate that this training procedure allows tackling challenging blind inverse problems.Our experiments include partial Fourier sampling problems arising in magnetic resonance imaging (MRI) with sensitivity estimation and off-resonance effects, computerized tomography (CT) with a tilted geometry and image deblurring with Fresnel diffraction kernels.

Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 37 of its 107 references that are in the map. First posted on arXiv 23 Feb 2022.

Builds on

37 of 107 references.

References · 37

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  3. 2017Deep Image Priorcited by 1,097
  4. 2020Plug-and-Play Image Restoration with Deep Denoiser Priorcited by 467
  5. 2018fastMRI: An Open Dataset and Benchmarks for Accelerated MRIcited by 464
  6. 2016Deep Convolutional Neural Network for Inverse Problems in Imagingcited by 431
  7. 2017MoDL: Model Based Deep Learning Architecture for Inverse Problemscited by 384
  8. 2017Learning a Variational Network for Reconstruction of Accelerated MRI Datacited by 340
  9. 2017Compressed Sensing using Generative Modelscited by 337
  10. 2017Learned Primal-dual Reconstructioncited by 251
  11. 2017Image reconstruction by domain transform manifold learningcited by 231
  12. 2019On instabilities of deep learning in image reconstruction - Does AI come at a cost?cited by 211
25 more
  1. 2019Blind Super-Resolution With Iterative Kernel Correctioncited by 203
  2. 2019Plug-and-Play Methods Provably Converge with Properly Trained Denoiserscited by 197
  3. 2015Learning a Convolutional Neural Network for Non-uniform Motion Blur Removalcited by 148
  4. 2017Solving ill-posed inverse problems using iterative deep neural networkscited by 145
  5. 2021Fit without fear: remarkable mathematical phenomena of deep learning through the prism of interpolationcited by 130
  6. 2021Bayesian imaging using Plug & Play priors: when Langevin meets Tweediecited by 96
  7. 2018Denoising Prior Driven Deep Neural Network for Image Restorationcited by 80
  8. 2016A Neural Approach to Blind Motion Deblurringcited by 69
  9. 2014Learning to Deblurcited by 66
  10. 2020Solving Inverse Problems With Deep Neural Networks -- Robustness Included?cited by 58
  11. 2020TorchRadon: Fast Differentiable Routines for Computed Tomographycited by 44
  12. 2017Unrolled Optimization with Deep Priorscited by 40
  13. 2018Blind Image Deconvolution using Deep Generative Priorscited by 40
  14. 2018Learning The Invisible: A Hybrid Deep Learning-Shearlet Framework for Limited Angle Computed Tomographycited by 29
  15. 2021B-spline Parameterized Joint Optimization of Reconstruction and K-space Trajectories (BJORK) for Accelerated 2D MRIcited by 29
  16. 2018Graph-Based Blind Image Deblurring From a Single Photographcited by 24
  17. 2019PILOT: Physics-Informed Learned Optimized Trajectories for Accelerated MRIcited by 24
  18. 2020Deep Phase Decoder: Self-calibrating phase microscopy with an untrained deep neural networkcited by 24
  19. 2021Lucas-Kanade Reloaded: End-to-End Super-Resolution from Raw Image Burstscited by 21
  20. 2020Model Adaptation for Inverse Problems in Imagingcited by 20
  21. 2019An Algorithm Unrolling Approach to Deep Image Deblurringcited by 14
  22. 2021Wasserstein Patch Prior for Image Superresolutioncited by 13
  23. 2021cuFINUFFT: a load-balanced GPU library for general-purpose nonuniform FFTscited by 11
  24. 2019Sigma-net: Systematic Evaluation of Iterative Deep Neural Networks for Fast Parallel MR Image Reconstructioncited by 7
  25. 2021Subtle Data Crimes: Naively training machine learning algorithms could lead to overly-optimistic resultscited by 4

Built on by

Cited by year

Citing papers per year: 2023: 1; 2024: 1; 2025: 1.

3 papers.

  1. 2023Block Coordinate Plug-and-Play Methods for Blind Inverse Problemscited by 8
  2. 2024Robustness and Exploration of Variational and Machine Learning Approaches to Inverse Problems: An Overviewcited by 1
  3. 2025Reconstruct Anything Model: a lightweight general model for computational imagingcited by 1