Learned Primal-dual Reconstruction

Abstract

We propose the Learned Primal-Dual algorithm for tomographic reconstruction. The algorithm accounts for a (possibly non-linear) forward operator in a deep neural network by unrolling a proximal primal-dual optimization method, but where the proximal operators have been replaced with convolutional neural networks. The algorithm is trained end-to-end, working directly from raw measured data and it does not depend on any initial reconstruction such as FBP. We compare performance of the proposed method on low dose CT reconstruction against FBP, TV, and deep learning based post-processing of FBP. For the Shepp-Logan phantom we obtain >6dB PSNR improvement against all compared methods. For human phantoms the corresponding improvement is 6.6dB over TV and 2.2dB over learned post-processing along with a substantial improvement in the SSIM. Finally, our algorithm involves only ten forward-back-projection computations, making the method feasible for time critical clinical applications.

11 pages, 5 figures

References in corpus (3)

• Low-Dose CT with a Residual Encoder-Decoder Convolutional Neural Network (RED-CNN)
• Solving ill-posed inverse problems using iterative deep neural networks
• Learning Proximal Operators: Using Denoising Networks for Regularizing Inverse Imaging Problems

Cited by in corpus (98)

• CT Super-resolution GAN Constrained by the Identical, Residual, and Cycle Learning Ensemble(GAN-CIRCLE)
• Model based learning for accelerated, limited-view 3D photoacoustic tomography
• DeepPET: A deep encoder-decoder network for directly solving the PET reconstruction inverse problem
• FISTA-Net: Learning A Fast Iterative Shrinkage Thresholding Network for Inverse Problems in Imaging
• Results of the 2020 fastMRI Challenge for Machine Learning MR Image Reconstruction
• AMP-Net: Denoising based Deep Unfolding for Compressive Image Sensing
• Image Reconstruction: From Sparsity to Data-adaptive Methods and Machine Learning
• Advancing machine learning for MR image reconstruction with an open competition: Overview of the 2019 fastMRI challenge
• Cycle Consistent Adversarial Denoising Network for Multiphase Coronary CT Angiography
• Adaptive Diffusion Priors for Accelerated MRI Reconstruction
• InversionNet: A Real-Time and Accurate Full Waveform Inversion with CNNs and continuous CRFs
• A Review on Deep Learning in Medical Image Reconstruction
• Computed Tomography Reconstruction Using Deep Image Prior and Learned Reconstruction Methods
• Deep Learning in Photoacoustic Tomography: Current approaches and future directions
• Solving Inverse Problems With Deep Neural Networks -- Robustness Included?
• Noise2Inverse: Self-supervised deep convolutional denoising for tomography
• Convolutional Sparse Coding for Compressed Sensing CT Reconstruction
• Learning The Invisible: A Hybrid Deep Learning-Shearlet Framework for Limited Angle Computed Tomography
• NAF: Neural Attenuation Fields for Sparse-View CBCT Reconstruction
• The LoDoPaB-CT Dataset: A Benchmark Dataset for Low-Dose CT Reconstruction Methods
• Dense Recurrent Neural Networks for Accelerated MRI: History-Cognizant Unrolling of Optimization Algorithms
• Spatio-Temporal Deep Learning-Based Undersampling Artefact Reduction for 2D Radial Cine MRI with Limited Data
• Regularization by architecture: A deep prior approach for inverse problems
• Deep learning in biomedical optics
• Computationally Efficient Deep Neural Network for Computed Tomography Image Reconstruction
• PYRO-NN: Python Reconstruction Operators in Neural Networks
• A Partially Learned Algorithm for Joint Photoacoustic Reconstruction and Segmentation
• Structural engineering from an inverse problems perspective
• Deep Learning Methods for Partial Differential Equations and Related Parameter Identification Problems
• MoDL-MUSSELS: Model-Based Deep Learning for Multi-Shot Sensitivity Encoded Diffusion MRI
• Deep Learning for space-variant deconvolution in galaxy surveys
• Low-Dose CT with Deep Learning Regularization via Proximal Forward Backward Splitting
• Physics-/Model-Based and Data-Driven Methods for Low-Dose Computed Tomography: A survey
• Deep learning reconstruction of digital breast tomosynthesis images for accurate breast density and patient-specific radiation dose estimation
• Task adapted reconstruction for inverse problems
• Neural Networks-based Regularization for Large-Scale Medical Image Reconstruction
• PARALLELPROJ -- An open-source framework for fast calculation of projections in tomography
• CDLNet: Noise-Adaptive Convolutional Dictionary Learning Network for Blind Denoising and Demosaicing
• Known Operator Learning and Hybrid Machine Learning in Medical Imaging -- A Review of the Past, the Present, and the Future
• Deep Unfolding with Normalizing Flow Priors for Inverse Problems
• SGD-Net: Efficient Model-Based Deep Learning with Theoretical Guarantees
• Systematic Review on Learning-based Spectral CT
• Multi-modal Deep Guided Filtering for Comprehensible Medical Image Processing
• Infinite-dimensional inverse problems with finite measurements
• Noise2Context: Context-assisted Learning 3D Thin-layer Low Dose CT Without Clean Data
• Unrolled Primal-Dual Networks for Lensless Cameras
• An Unsupervised Reconstruction Method For Low-Dose CT Using Deep Generative Regularization Prior
• A study of why we need to reassess full reference image quality assessment with medical images
• PatchNR: Learning from Very Few Images by Patch Normalizing Flow Regularization
• SPULTRA: Low-Dose CT Image Reconstruction with Joint Statistical and Learned Image Models
• Rotation Equivariant Proximal Operator for Deep Unfolding Methods in Image Restoration
• Data-driven nonsmooth optimization
• Unified Supervised-Unsupervised (SUPER) Learning for X-ray CT Image Reconstruction
• ENSURE: A General Approach for Unsupervised Training of Deep Image Reconstruction Algorithms
• Score-Based Generative Models for PET Image Reconstruction
• DIR-DBTnet: Deep iterative reconstruction network for 3D digital breast tomosynthesis imaging
• Metappearance: Meta-Learning for Visual Appearance Reproduction
• Unsupervised Knowledge-Transfer for Learned Image Reconstruction
• Assessment of Data Consistency through Cascades of Independently Recurrent Inference Machines for fast and robust accelerated MRI reconstruction
• Deep Learning for Biomedical Image Reconstruction: A Survey
• Inverse problems on low-dimensional manifolds
• PtychoDV: Vision Transformer-Based Deep Unrolling Network for Ptychographic Image Reconstruction
• Phase Unwrapping of Color Doppler Echocardiography using Deep Learning
• Nest-DGIL: Nesterov-optimized Deep Geometric Incremental Learning for CS Image Reconstruction
• Recurrent Localization Networks applied to the Lippmann-Schwinger Equation
• Efficient Physics-Based Learned Reconstruction Methods for Real-Time 3D Near-Field MIMO Radar Imaging
• Spatiotemporal PET reconstruction using ML-EM with learned diffeomorphic deformation
• Learning Geometry-Dependent and Physics-Based Inverse Image Reconstruction
• A Compound Gaussian Least Squares Algorithm and Unrolled Network for Linear Inverse Problems
• NESTANets: Stable, accurate and efficient neural networks for analysis-sparse inverse problems
• Semi-Blind Source Separation with Learned Constraints
• End-to-end Memory-Efficient Reconstruction for Cone Beam CT
• Learning Task-Specific Strategies for Accelerated MRI
• A plug-and-play framework for curvilinear structure segmentation based on a learned reconnecting regularization
• A signal detection model for quantifying over-regularization in non-linear image reconstruction
• Learned Interferometric Imaging for the SPIDER Instrument
• Learned Cone-Beam CT Reconstruction Using Neural Ordinary Differential Equations
• A Dataset-free Deep learning Method for Low-Dose CT Image Reconstruction
• Accelerating innovation with software abstractions for scalable computational geophysics
• Deep network series for large-scale high-dynamic range imaging
• Invertible residual networks in the context of regularization theory for linear inverse problems
• Deep learning-based deconvolution for interferometric radio transient reconstruction
• NoSENSE: Learned unrolled cardiac MRI reconstruction without explicit sensitivity maps
• Deep Unrolled Network for Video Super-Resolution
• Cutting Voxel Projector a New Approach to Construct 3D Cone Beam CT Operator
• Learned Primal Dual Splitting for Self-Supervised Noise-Adaptive MRI Reconstruction
• Phase Retrieval using Expectation Consistent Signal Recovery Algorithm based on Hypernetwork
• A Statistical Framework to Investigate the Optimality of Signal-Reconstruction Methods
• Deep learning based dictionary learning and tomographic image reconstruction
• Equivariant neural networks for inverse problems
• Compressing Sign Information in DCT-based Image Coding via Deep Sign Retrieval
• Swap-Net: A Memory-Efficient 2.5D Network for Sparse-View 3D Cone Beam CT Reconstruction
• Deep Unfolding-Aided Parameter Tuning for Plug-and-Play-Based Video Snapshot Compressive Imaging
• Sparse View Tomographic Reconstruction of Elongated Objects using Learned Primal-Dual Networks
• CPSNet: Physics-Inspired Label-Free Deep Unfolding for Lung Ultrasound B-Line Detection
• Equivariance2Inverse: A Practical Self-Supervised CT Reconstruction Method Benchmarked on Real, Limited-Angle, and Blurred Data
• Deep Inertia $L_p$ Half-Quadratic Splitting Unrolling Network for Sparse View CT Reconstruction
• Deep learning for inverse problems with unknown operator