Leveraging the Structure of Medical Data for Improved Representation Learning

TL;DR

This paper introduces a self-supervised learning framework that exploits the inherent multi-view structure of medical imaging data, specifically chest X-rays, to improve representation learning without textual labels.

Contribution

It presents a novel, structure-aware pretraining method for medical images that enhances data efficiency and domain-specific feature extraction.

Findings

Outperforms supervised methods on MIMIC-CXR

Requires no textual annotations for training

Leverages multi-view structure for better representations

Abstract

Building generalizable medical AI systems requires pretraining strategies that are data-efficient and domain-aware. Unlike internet-scale corpora, clinical datasets such as MIMIC-CXR offer limited image counts and scarce annotations, but exhibit rich internal structure through multi-view imaging. We propose a self-supervised framework that leverages the inherent structure of medical datasets. Specifically, we treat paired chest X-rays (i.e., frontal and lateral views) as natural positive pairs, learning to reconstruct each view from sparse patches while aligning their latent embeddings. Our method requires no textual supervision and produces informative representations. Evaluated on MIMIC-CXR, we show strong performance compared to supervised objectives and baselines being trained without leveraging structure. This work provides a lightweight, modality-agnostic blueprint for domain-specific pretraining where data is structured but scarce