Improving Joint Learning of Chest X-Ray and Radiology Report by Word Region Alignment

TL;DR

This paper introduces JoImTeRNet, a self-supervised joint learning framework for chest X-ray images and radiology reports, leveraging global and local visual-textual alignment to improve downstream medical imaging tasks.

Contribution

The paper presents a novel multi-modal pre-training method that aligns image regions with report words using attention, enhancing representation learning without manual supervision.

Findings

Improved cross-modality retrieval performance

Enhanced multi-label classification accuracy

Effective local and global alignment of image and text features

Abstract

Self-supervised learning provides an opportunity to explore unlabeled chest X-rays and their associated free-text reports accumulated in clinical routine without manual supervision. This paper proposes a Joint Image Text Representation Learning Network (JoImTeRNet) for pre-training on chest X-ray images and their radiology reports. The model was pre-trained on both the global image-sentence level and the local image region-word level for visual-textual matching. Both are bidirectionally constrained on Cross-Entropy based and ranking-based Triplet Matching Losses. The region-word matching is calculated using the attention mechanism without direct supervision about their mapping. The pre-trained multi-modal representation learning paves the way for downstream tasks concerning image and/or text encoding. We demonstrate the representation learning quality by cross-modality retrievals and multi-label classifications on two datasets: OpenI-IU and MIMIC-CXR