Domain Adaptation to move from 1.5T to 3T MRI

Integrating 3D magnetic resonance imaging (MRI) with machine learning has shown promising results in healthcare, especially in detecting Alzheimer's Disease (AD). However, changes in MRI technologies and acquisition protocols often yield limited data, leading to potential overfitting. From the machine learning perspective, data acquired with 1.5T follow a different distribution w.r.t. those acquired with 3T MRIs.

The goal of the project is to investigate transfer learning techniques and how to adapt the models obtained from historical data to address changes in MRI acquisitions (from 1.5T to 3T MRI). 

Furthermore, the study considers the real-case scenario in which new and few 3T MRI scans have been acquired, but historical data is unavailable. In this context, is it possible to rely on 2D models pre-trained on natural images (e.g., ImageNet)?