Deep Convolutional Neural Network for Inverse Problems in Imaging
From MaRDI portal
Abstract: In this paper, we propose a novel deep convolutional neural network (CNN)-based algorithm for solving ill-posed inverse problems. Regularized iterative algorithms have emerged as the standard approach to ill-posed inverse problems in the past few decades. These methods produce excellent results, but can be challenging to deploy in practice due to factors including the high computational cost of the forward and adjoint operators and the difficulty of hyper parameter selection. The starting point of our work is the observation that unrolled iterative methods have the form of a CNN (filtering followed by point-wise non-linearity) when the normal operator (H*H, the adjoint of H times H) of the forward model is a convolution. Based on this observation, we propose using direct inversion followed by a CNN to solve normal-convolutional inverse problems. The direct inversion encapsulates the physical model of the system, but leads to artifacts when the problem is ill-posed; the CNN combines multiresolution decomposition and residual learning in order to learn to remove these artifacts while preserving image structure. We demonstrate the performance of the proposed network in sparse-view reconstruction (down to 50 views) on parallel beam X-ray computed tomography in synthetic phantoms as well as in real experimental sinograms. The proposed network outperforms total variation-regularized iterative reconstruction for the more realistic phantoms and requires less than a second to reconstruct a 512 x 512 image on GPU.
Cited in
(only showing first 100 items - show all)- A dataset-free deep learning method for low-dose CT image reconstruction
- Adversarially learned iterative reconstruction for imaging inverse problems
- Networks for nonlinear diffusion problems in imaging
- Regularization by architecture: a deep prior approach for inverse problems
- Scaling Up Bayesian Uncertainty Quantification for Inverse Problems Using Deep Neural Networks
- Exploring structural sparsity of coil images from 3-dimensional directional tight framelets for SENSE reconstruction
- NF-ULA: normalizing flow-based unadjusted Langevin algorithm for imaging inverse problems
- Regularization of inverse problems: deep equilibrium models versus bilevel learning
- Bilevel optimization, deep learning and fractional Laplacian regularization with applications in tomography
- Constrained and unconstrained deep image prior optimization models with automatic regularization
- A data-dependent regularization method based on the graph Laplacian
- Determination of singular value truncation threshold for regularization in ill-posed problems
- Attenuating and enhancing properties of the approximate deconvolution method based on higher-order explicit and compact filters
- A Note on the Regularity of Images Generated by Convolutional Neural Networks
- Mini-workshop: Deep learning and inverse problems. Abstracts from the mini-workshop held March 4--10, 2018
- Strong Solutions for PDE-Based Tomography by Unsupervised Learning
- Lama-Net: a convergent network architecture for dual-domain reconstruction
- Deep learning architectures for nonlinear operator functions and nonlinear inverse problems
- What Kinds of Functions Do Deep Neural Networks Learn? Insights from Variational Spline Theory
- First arrival traveltime tomography using supervised descent learning technique
- Computerized tomography and reproducing kernels
- DRIP: deep regularizers for inverse problems
- Reversible Autoencoder: A CNN-Based Nonlinear Lifting Scheme for Image Reconstruction
- Training Very Deep CNNs for General Non-Blind Deconvolution
- Measuring Complexity of Learning Schemes Using Hessian-Schatten Total Variation
- Convergent data-driven regularizations for CT reconstruction
- Convolutional Neural Networks in Phase Space and Inverse Problems
- MiPhDUO: microwave imaging via physics-informed deep unrolled optimization
- A nonlocal image denoising algorithm using the structural similarity metric
- Deep convolutional framelets: a general deep learning framework for inverse problems
- On the randomised stability constant for inverse problems
- Analytic continuation of noisy data using Adams Bashforth residual neural network
- A Helmholtz equation solver using unsupervised learning: application to transcranial ultrasound
- Recent trends and advances in solving the inverse problem for EEG source localization
- Training adaptive reconstruction networks for blind inverse problems
- A mathematical framework for deep learning in elastic source imaging
- Bilevel Methods for Image Reconstruction
- Deep neural networks for inverse problems with pseudodifferential operators: an application to limited-angle tomography
- Robustness of data-driven approaches in limited angle tomography
- Non-invasive inference of thrombus material properties with physics-informed neural networks
- Wasserstein-based projections with applications to inverse problems
- A new overhead crane emergency braking method with theoretical analysis and experimental verification
- Deep Learning--Based Dictionary Learning and Tomographic Image Reconstruction
- Image restoration based on transformed total variation and deep image prior
- On Learning the Invisible in Photoacoustic Tomography with Flat Directionally Sensitive Detector
- Feasibility-based fixed point networks
- Learning the sparse prior: modern approaches
- Learned iterative reconstruction
- Learned regularizers for inverse problems
- Regularization of inverse problems by neural networks
- Shearlets: from theory to deep learning
- Provably convergent learned inexact descent algorithm for low-dose CT reconstruction
- An over complete deep learning method for inverse problems
- Equivariant neural networks for inverse problems
- Deep learning based source reconstruction method using asymmetric encoder-decoder structure and physics-induced loss
- Neural network approximation of continuous functions in high dimensions with applications to inverse problems
- Sampling type method combined with deep learning for inverse scattering with one incident wave
- Improvements on uncertainty quantification with variational autoencoders
- Application of machine learning regression models to inverse eigenvalue problems
- Degradation simulator for infinite horizon controlled linear time-invariant systems
- Sequential Bayesian design for efficient surrogate construction in the inversion of Darcy flows
- Neural networks for Bayesian inverse problems governed by a nonlinear ODE
- Shallow neural networks for fluid flow reconstruction with limited sensors
- Multi-layer 5D optical data storage: mathematical modeling and deep learning-based reconstruction of birefringent parameters
- An introduction to the mathematics of deep learning
- A data and knowledge driven approach for SPECT using convolutional neural networks and iterative algorithms
- Numerical methods for identifying the diffusion coefficient in a nonlinear elliptic equation
- The mathematics of artificial intelligence
- Maximum discrepancy generative regularization and non-negative matrix factorization for single channel source separation
- Deep learning for photoacoustic tomography from sparse data
- Deep learning on image denoising: an overview
- Bayesian inference for PDE-based inverse problems using the optimization of a discrete loss
- Unsupervised knowledge-transfer for learned image reconstruction
- Deep synthesis network for regularizing inverse problems
- A primal-dual algorithm for image reconstruction with input-convex neural network regularizers
- Normalizing flow regularization for photoacoustic tomography
- Electrical impedance tomography with deep Calderón method
- Learning the invisible: a hybrid deep learning-shearlet framework for limited angle computed tomography
- Instability in deep learning -- when algorithms cannot compute uncertainty quantifications for neural networks
- On finite-horizon \(H_\infty\) state estimation for discrete-time delayed memristive neural networks under stochastic communication protocol
- WARPd: a linearly convergent first-order primal-dual algorithm for inverse problems with approximate sharpness conditions
- A weaker adversarial neural networks method for linear second order elliptic PDEs
- Convergent regularization in inverse problems and linear plug-and-play denoisers
- Learn an index operator by CNN for solving diffusive optical tomography: a deep direct sampling method
- A blind Poisson-Gaussian noise separation using learning copula densities
- Inverse problems are solvable on real number signal processing hardware
- Computed tomography reconstruction using deep image prior and learned reconstruction methods
- On asymptotic behavior for a class of diffusion equations involving the fractional \(\wp (\cdot)\)-Laplacian as \(\wp (\cdot)\) goes to \(\infty\)
- Learning to scan: a deep reinforcement learning approach for personalized scanning in CT imaging
- A review on deep learning in medical image reconstruction
- scientific article; zbMATH DE number 7625169 (Why is no real title available?)
- Stability bounds for the unfolded forward-backward algorithm
- Practical operator sketching framework for accelerating iterative data-driven solutions in linear inverse problems
- Invertible residual networks in the context of regularization theory for linear inverse problems
- TomoSelfDEQ: self-supervised deep equilibrium learning for sparse-angle CT reconstruction
- Neural-network-based regularization methods for inverse problems in imaging
- Data driven regularization by projection
- Learnable Douglas-Rachford iteration and its applications in DOT imaging
- A novel physics-informed framework for reconstruction of structural defects
- Data-consistent neural networks for solving nonlinear inverse problems
This page was built for publication: Deep Convolutional Neural Network for Inverse Problems in Imaging
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4619040)
