Towards Scalable Foundation Models for Digital Dermatology

TL;DR

This paper develops and evaluates domain-specific foundation models for digital dermatology using self-supervised learning on a large dataset, demonstrating superior performance over general models and enabling resource-efficient clinical applications.

Contribution

It introduces dermatology-specific foundation models trained with SSL on a large dataset, emphasizing smaller, resource-efficient models for clinical use, and compares various SSL methods against existing models.

Findings

Models outperform general-purpose models on dermatology tasks.

Pre-trained models approach the performance of much larger models.

Public release of training code and models to aid clinical research.

Abstract

The growing demand for accurate and equitable AI models in digital dermatology faces a significant challenge: the lack of diverse, high-quality labeled data. In this work, we investigate the potential of domain-specific foundation models for dermatology in addressing this challenge. We utilize self-supervised learning (SSL) techniques to pre-train models on a dataset of over 240,000 dermatological images from public and private collections. Our study considers several SSL methods and compares the resulting foundation models against domain-agnostic models like those pre-trained on ImageNet and state-of-the-art models such as MONET across 12 downstream tasks. Unlike previous research, we emphasize the development of smaller models that are more suitable for resource-limited clinical settings, facilitating easier adaptation to a broad range of use cases. Results show that models pre-trained in this work not only outperform general-purpose models but also approach the performance of models 50 times larger on clinically relevant diagnostic tasks. To promote further research in this direction, we publicly release both the training code and the foundation models, which can benefit clinicians in dermatological applications.