Generalized Deep Learning-based Proximal Gradient Descent for MR Reconstruction

TL;DR

This paper introduces a generalized deep learning-based proximal gradient descent method for MR reconstruction that uses a model-independent regularization, improving generalization across different acquisition settings and undersampling patterns.

Contribution

The proposed method employs a pre-trained regularization network that is independent of the forward model, enhancing adaptability in various MR imaging scenarios.

Findings

Achieved approximately 3 dB improvement in PSNR over conventional L1 regularization.

Demonstrated robustness across different MR acquisition settings.

Showed flexibility in handling various undersampling patterns.

Abstract

The data consistency for the physical forward model is crucial in inverse problems, especially in MR imaging reconstruction. The standard way is to unroll an iterative algorithm into a neural network with a forward model embedded. The forward model always changes in clinical practice, so the learning component's entanglement with the forward model makes the reconstruction hard to generalize. The deep learning-based proximal gradient descent was proposed and use a network as regularization term that is independent of the forward model, which makes it more generalizable for different MR acquisition settings. This one-time pre-trained regularization is applied to different MR acquisition settings and was compared to conventional L1 regularization showing ~3 dB improvement in the peak signal-to-noise ratio. We also demonstrated the flexibility of the proposed method in choosing different undersampling patterns.