Learning spatially adaptive sparsity level maps for arbitrary convolutional dictionaries

TL;DR

This paper introduces an improved, interpretable image reconstruction method that adaptively learns sparsity maps for convolutional dictionaries, enhancing robustness and flexibility in MRI applications.

Contribution

It extends a model-based reconstruction approach with neural network-inferred adaptive sparsity maps, achieving filter-permutation invariance and enabling dictionary changes at inference.

Findings

Demonstrates improved robustness to data distribution shifts.

Shows benefits of using different dictionaries in MRI reconstruction.

Achieves filter-permutation invariance in the reconstruction method.

Abstract

State-of-the-art learned reconstruction methods often rely on black-box modules that, despite their strong performance, raise questions about their interpretability and robustness. Here, we build on a recently proposed image reconstruction method, which is based on embedding data-driven information into a model-based convolutional dictionary regularization via neural network-inferred spatially adaptive sparsity level maps. By means of improved network design and dedicated training strategies, we extend the method to achieve filter-permutation invariance as well as the possibility to change the convolutional dictionary at inference time. We apply our method to low-field MRI and compare it to several other recent deep learning-based methods, also on in vivo data, where the benefit of using a different dictionary is demonstrated. We further assess the method's robustness when tested on in- and out-of-distribution data. When tested on the latter, the proposed method suffers less from the data distribution shift compared to the other learned methods, which we attribute to its reduced reliance on training data due to its underlying model-based reconstruction component.