Improving Structural MRI Preprocessing with Hybrid Transformer GANs

Abstract

Magnetic resonance imaging (MRI) is a technique that is widely used in practice to evaluate any pathologies in the human body. One of the areas of interest is the human brain. Naturally, MR images are low-resolution and contain noise due to signal interference, the patient’s body’s radio-frequency emissions and smaller Tesla coil counts in the machinery. There is a need to solve this problem, as MR tomographs that have the capability of capturing high-resolution images are extremely expensive and the length of the procedure to capture such images increases by the order of magnitude. Vision transformers have lately shown state-of-the-art results in super-resolution tasks; therefore, we decided to evaluate whether we can employ them for structural MRI super-resolution tasks. A literature review showed that similar methods do not focus on perceptual image quality because upscaled images are often blurry and are subjectively of poor quality. Knowing this, we propose a methodology called HR-MRI-GAN, which is a hybrid transformer generative adversarial network capable of increasing resolution and removing noise from 2D T1w MRI slice images. Experiments show that our method quantitatively outperforms other SOTA methods in terms of perceptual image quality and is capable of subjectively generalizing to unseen data. During the experiments, we additionally identified that the visual saliency-induced index metric is not applicable to MRI perceptual quality assessment and that general-purpose denoising networks are effective when removing noise from MR images.

Speeding Up and Improving Image Quality in Glioblastoma MRI Protocol by Deep Learning Image Reconstruction

“…The use of DL for MRI reconstructions has the potential to reduce the examination time for TSE and FLAIR acquisitions. In recent years, a number of different approaches to deep learning (DL) have emerged, and a growing number of new developments in artificial networks have initiated a paradigm shift in the field of medical imaging [23][24][25][26].…”

Intelligent MRI diagnosis of neurological alterations in infants from 4 to 12 months

“…Also, with the increasing development of generative algorithms such as generative adversarial networks (GAN), we will seek to improve the quality because techniques such as super-resolution applied to magnetic resonance images have a positive impact on the interpretation by medical professionals, and classification models, facilitating the identification and classification of different brain abnormalities. 75…”

Positive Effect of Super-Resolved Structural Magnetic Resonance Imaging for Mild Cognitive Impairment Detection

“…For the super-resolution model improvements, we used the same ultra-high-resolution MRI dataset "human phantom" [52] that we used in our previous work [25]. (Dataset available online: https://datadryad.org/stash/dataset/doi:10.5061/dryad.38s74-accessed on 5 March 2024).…”

“…The baseline of the improvements for this study is our previously published method [25], which increases the resolution of structural MRIs while preserving perceptional image quality. It uses hybrid attention transformer (HAT) as a generator and introduces an adversarial training pipeline, which allows one to super-resolve structural MRI and decrease its blurriness and noise.…”

“…In the medical imaging field, preserving the structural part of the image quality is as important as the perceptual part. Therefore, just like in our previous paper [25], we maintain the focus to improve the main important aspects of the image quality-structural and perceptual.…”