Abstract:SIAM Journal on Imaging Sciences, Volume 16, Issue 4, Page 2202-2246, December 2023. We introduce a method for the fast estimation of data-adapted, spatially and temporally dependent regularization parameter-maps for variational image reconstruction, focusing on total variation (TV) minimization. The proposed approach is inspired by recent developments in algorithm unrolling using deep neural networks (NNs) and relies on two distinct subnetworks. The first subnetwork estimates the regularization parameter-map from the input data. The second subnetwork unrolls [math] iterations of an iterative algorithm which approximately solves the corresponding TV-minimization problem incorporating the previously estimated regularization parameter-map. The overall network is then trained end-to-end in a supervised learning fashion using pairs of clean and corrupted data but crucially without the need for access to labels for the optimal regularization parameter-maps. We first prove consistency of the unrolled scheme by showing that the unrolled minimizing energy functional used for the supervised learning [math]-converges, as [math] tends to infinity, to the corresponding functional that incorporates the exact solution map of the TV-minimization problem. Then, we apply and evaluate the proposed method on a variety of large-scale and dynamic imaging problems with retrospectively simulated measurement data for which the automatic computation of such regularization parameters has been so far challenging using the state-of-the-art methods: a 2D dynamic cardiac magnetic resonance imaging (MRI) reconstruction problem, a quantitative brain MRI reconstruction problem, a low-dose computed tomography problem, and a dynamic image denoising problem. The proposed method consistently improves the TV reconstructions using scalar regularization parameters, and the obtained regularization parameter-maps adapt well to imaging problems and data by leading to the preservation of detailed features. Although the choice of the regularization parameter-maps is data-driven and based on NNs, the subsequent reconstruction algorithm is interpretable since it inherits the properties (e.g., convergence guarantees) of the iterative reconstruction method from which the network is implicitly defined.

Regularization with Sparse Vector Fields: From Image Compression to TV-type Reconstruction

Efficient Compressed Sensing Reconstruction Using Group Sparse Total Variation Regularization

Image Reconstruction from Incomplete Convolution Data Via Total Variation Regularization

High-Accuracy Total Variation for Compressed Video Sensing

Compressive total variation for image reconstruction and restoration

Image Restoration by Combined Order Regularization with Optimal Spatial Adaptation

Supervised Learning of Sparsity-Promoting Regularizers for Denoising

Non-convex regularization based on shrinkage penalty function

A New Alternating Minimization Algorithm for Total Variation Image Reconstruction

Image Restoration using Total Variation Regularized Deep Image Prior

Structurally Adaptive Multi-Derivative Regularization for Image Recovery from Sparse Fourier Samples

Variational Bayes image restoration with compressive autoencoders

Compressed Sensing MRI Reconstruction Regularized by VAEs with Structured Image Covariance

Learning Regularization Parameter-Maps for Variational Image Reconstruction Using Deep Neural Networks and Algorithm Unrolling

Tensor Compressed Video Sensing Reconstruction by Combination of Fractional-Order Total Variation and Sparsifying Transform

Geometric properties of solutions to the total variation denoising problem

Haar Wavelets, Gradients and Approximate TV Regularization

A universal variational framework for sparsity-based image inpainting

A bi-level view of inpainting - based image compression

Compressed Sensing Parallel MRI with Adaptive Shrinkage TV Regularization

Graph-Based Non-Convex Low-Rank Regularization for Image Compression Artifact Reduction.