Cross Modal Global Local Representation Learning from Radiology Reports and X-Ray Chest Images

TL;DR

This paper introduces a multimodal representation learning approach combining radiology reports and X-ray images using attention mechanisms, achieving high accuracy in diagnosing lung pathologies with publicly available data.

Contribution

It presents a novel local and global multimodal representation learning method for radiology, utilizing attention mechanisms and publicly available datasets for lung pathology diagnosis.

Findings

Average AUC for pathology classification ranged from 0.85 to 0.87.

The method outperforms previous studies on IU-RR dataset.

Consistent results in lung pathology classification using multimodal representations.

Abstract

Deep learning models can be applied successfully in real-work problems; however, training most of these models requires massive data. Recent methods use language and vision, but unfortunately, they rely on datasets that are not usually publicly available. Here we pave the way for further research in the multimodal language-vision domain for radiology. In this paper, we train a representation learning method that uses local and global representations of the language and vision through an attention mechanism and based on the publicly available Indiana University Radiology Report (IU-RR) dataset. Furthermore, we use the learned representations to diagnose five lung pathologies: atelectasis, cardiomegaly, edema, pleural effusion, and consolidation. Finally, we use both supervised and zero-shot classifications to extensively analyze the performance of the representation learning on the IU-RR dataset. Average Area Under the Curve (AUC) is used to evaluate the accuracy of the classifiers for classifying the five lung pathologies. The average AUC for classifying the five lung pathologies on the IU-RR test set ranged from 0.85 to 0.87 using the different training datasets, namely CheXpert and CheXphoto. These results compare favorably to other studies using UI-RR. Extensive experiments confirm consistent results for classifying lung pathologies using the multimodal global local representations of language and vision information.