Noise2Recon: Enabling Joint MRI Reconstruction and Denoising with Semi-Supervised and Self-Supervised Learning

TL;DR

Noise2Recon is a semi-supervised and self-supervised learning method that jointly performs MRI reconstruction and denoising, effectively handling limited labeled data and out-of-distribution noise, outperforming traditional and supervised approaches.

Contribution

It introduces a model-agnostic consistency training framework for joint MRI reconstruction and denoising that works with both labeled and unlabeled data, improving robustness and reducing data requirements.

Findings

Outperforms baselines with limited or no labeled data.

Matches supervised models trained with 14x more data.

Excels in low-SNR and out-of-distribution scenarios.

Abstract

Deep learning (DL) has shown promise for faster, high quality accelerated MRI reconstruction. However, supervised DL methods depend on extensive amounts of fully-sampled (labeled) data and are sensitive to out-of-distribution (OOD) shifts, particularly low signal-to-noise ratio (SNR) acquisitions. To alleviate this challenge, we propose Noise2Recon, a model-agnostic, consistency training method for joint MRI reconstruction and denoising that can use both fully-sampled (labeled) and undersampled (unlabeled) scans in semi-supervised and self-supervised settings. With limited or no labeled training data, Noise2Recon outperforms compressed sensing and deep learning baselines, including supervised networks, augmentation-based training, fine-tuned denoisers, and self-supervised methods, and matches performance of supervised models, which were trained with 14x more fully-sampled scans. Noise2Recon also outperforms all baselines, including state-of-the-art fine-tuning and augmentation techniques, among low-SNR scans and when generalizing to other OOD factors, such as changes in acceleration factors and different datasets. Augmentation extent and loss weighting hyperparameters had negligible impact on Noise2Recon compared to supervised methods, which may indicate increased training stability. Our code is available at https://github.com/ad12/meddlr.