An open-source framework for cloud data processing from low-field scanners

Motivation: Advanced data processing tasks, such as model-based image reconstructions, require computational resources unavailable in many affordable and portable MRI systems. Goal(s): The main goal is to release the first open-source framework for cloud data processing, compatible with the MaRCoS and MaRGE ecosystem developed for low-field MRI scanners. Approach: We have upgraded MaRGE to integrate with Tyger, an open-source framework developed by Microsoft Research for remote signal processing. Results: We have demonstrated a successful integration of Tyger with MaRGE, enabling standard and advanced image reconstruction in the cloud, and demonstrating also compatibility with a deep learning network. Impact: This work connects affordable low-field MRI scanners to some of the most powerful computational cloud resources, where it shall be possible to execute strictly any tool that can be loaded into a Docker image.