Disease-informed Adaptation of Vision-Language Models

TL;DR

This paper proposes a disease-informed adaptation method for vision-language models that improves their ability to recognize new and underrepresented diseases in medical imaging, even with limited data.

Contribution

It introduces a novel disease prototype learning framework utilizing contextual prompting to enhance VLMs' disease concept understanding.

Findings

Significant performance improvements over existing methods.

Effective recognition of new and rare diseases with limited data.

Enhanced transfer learning capabilities in medical image analysis.

Abstract

In medical image analysis, the expertise scarcity and the high cost of data annotation limits the development of large artificial intelligence models. This paper investigates the potential of transfer learning with pre-trained vision-language models (VLMs) in this domain. Currently, VLMs still struggle to transfer to the underrepresented diseases with minimal presence and new diseases entirely absent from the pretraining dataset. We argue that effective adaptation of VLMs hinges on the nuanced representation learning of disease concepts. By capitalizing on the joint visual-linguistic capabilities of VLMs, we introduce disease-informed contextual prompting in a novel disease prototype learning framework. This approach enables VLMs to grasp the concepts of new disease effectively and efficiently, even with limited data. Extensive experiments across multiple image modalities showcase notable enhancements in performance compared to existing techniques.