Learn to Ignore: Domain Adaptation for Multi-Site MRI Analysis

TL;DR

This paper introduces a novel domain adaptation method called Learn to Ignore (L2I) that improves multi-site MRI classification by reducing scanner-related biases in the data.

Contribution

The paper proposes a new approach to ignore scanner-specific features in MR images through constraints on the latent space, enhancing classification performance across sites.

Findings

L2I outperforms existing domain adaptation methods on multi-site MRI data.

The method effectively reduces scanner bias in MR image classification.

Improves generalization in small datasets with multi-site variability.

Abstract

The limited availability of large image datasets, mainly due to data privacy and differences in acquisition protocols or hardware, is a significant issue in the development of accurate and generalizable machine learning methods in medicine. This is especially the case for Magnetic Resonance (MR) images, where different MR scanners introduce a bias that limits the performance of a machine learning model. We present a novel method that learns to ignore the scanner-related features present in MR images, by introducing specific additional constraints on the latent space. We focus on a real-world classification scenario, where only a small dataset provides images of all classes. Our method \textit{Learn to Ignore (L2I)} outperforms state-of-the-art domain adaptation methods on a multi-site MR dataset for a classification task between multiple sclerosis patients and healthy controls.