Simultaneous Image Quality Improvement and Artefacts Correction in Accelerated MRI

TL;DR

The paper tackles the challenge of speeding up MRI scans without compromising diagnostic quality by simultaneously addressing under-sampling and artefacts. It introduces USArt, a dual-domain network with K-net performing k-space completion and I-net refining image-domain artefacts, trained sequentially with MS-SSIM-based losses. Evaluations on 2D Cartesian mouse brain data show that a 5x acceleration is feasible with minimal quality loss and that USArt remains robust to Gaussian noise and motion artefacts, outperforming a prior baseline. The approach promises faster, more reliable MRI reconstructions suitable for real-world clinical scenarios and lays groundwork for extensions to other trajectories and architectures.

Abstract

MR data are acquired in the frequency domain, known as k-space. Acquiring high-quality and high-resolution MR images can be time-consuming, posing a significant challenge when multiple sequences providing complementary contrast information are needed or when the patient is unable to remain in the scanner for an extended period of time. Reducing k-space measurements is a strategy to speed up acquisition, but often leads to reduced quality in reconstructed images. Additionally, in real-world MRI, both under-sampled and full-sampled images are prone to artefacts, and correcting these artefacts is crucial for maintaining diagnostic accuracy. Deep learning methods have been proposed to restore image quality from under-sampled data, while others focused on the correction of artefacts that result from the noise or motion. No approach has however been proposed so far that addresses both acceleration and artefacts correction, limiting the performance of these models when these degradation factors occur simultaneously. To address this gap, we present a method for recovering high-quality images from under-sampled data with simultaneously correction for noise and motion artefact called USArt (Under-Sampling and Artifact correction model). Customized for 2D brain anatomical images acquired with Cartesian sampling, USArt employs a dual sub-model approach. The results demonstrate remarkable increase of signal-to-noise ratio (SNR) and contrast in the images restored. Various under-sampling strategies and degradation levels were explored, with the gradient under-sampling strategy yielding the best outcomes. We achieved up to 5x acceleration and simultaneously artefacts correction without significant degradation, showcasing the model's robustness in real-world settings.

Figures (4)

• Figure 1: The preprocessing pipeline and USArt. A) Artifacts and noise are added to full k-space, before under-sampling is performed using specific masks and acceleration factors. This degraded k-space dataset is used as input for the USArt model. B) USArt utilizes two U-Net based components: K-net and I-net. K-net operates in the k-space domain to fill missing lines, and its output is transformed to the image domain via an inverse Fourier Transform. I-net then refines this output, focusing on artifacts correction and image consistency.
• Figure 2: Under-sampling strategies with acceleration factor $5\times$: Left column from top to bottom) k-space, reconstructed, and zoom images (blue frame) for the original image. Grey panel) Under-sampled k-space and the zoom corresponding reconstructed images for gradient, random and uniform under-sampling. Blue-light panel) the corresponding USArt's output.
• Figure 3: Different acceleration factors: Left column) from top to bottom: k-space, reconstructed, and zoom images (blue frame) from the original image. Grey panel) Under-sampled k-space and the zoom corresponding reconstructed images for acceleration factors $2\times$, $5\times$, and $10\times$ . Blue panel) the corresponding USArt's output.
• Figure 4: Illustration of artefacts correction in accelerated images: First column) Original and zoom images (blue frame). Grey box) $5\times$under-sampled full and zoomed images with no artefacts, noise, motion artefact and their combination. Blue box) Corresponding images with quality restored by USArt and KIKI model's. US: Under-sampled, MA: Motion Artifact.