An open-source platform for image simulation, denoising, and super-resolution for point-of-care MRI devices

Motivation: Low-field MRI is increasingly being applied in lower- and middle-income countries, but due to limited resources, training is scare. Goal(s): Provide an open-source plateform for simulation, teaching, and a denoising and super-resolution pieline of low-field MRI images. Approach: Using 3D maps of relaxation times, proton density, B0 and low-field system-specific parameters, such as limited gradient linearity, simulation of low-field images are created. Application of bm4d denoising and AI super-resolution on low-field images is additionally proposed. Results: We provide an open-source graphical interface that can simulate and generate multiple sequences of low-field MRI, and a denoised and super-resolution pipeline increasing low-field image enhancement. Impact: This study provides a simple open-source python platform to simulate point-of-care low-field MRI images, reflecting specific system-specific parameters for teaching purposes, and a fast advanced denoising and AI-based super-resolution pipeline for low-field images.