MultiMedEval: A Benchmark and a Toolkit for Evaluating Medical Vision-Language Models

TL;DR

MultiMedEval is an open-source toolkit designed to facilitate fair, comprehensive, and reproducible evaluation of medical vision-language models across diverse tasks, datasets, and medical domains.

Contribution

It introduces a standardized, easy-to-use benchmarking toolkit and a broad evaluation framework for medical VLMs, promoting consistency and transparency in model assessment.

Findings

Evaluates models on 6 tasks across 23 datasets and 11 medical domains.

Provides a simple Python interface for quick model assessment.

Aims to standardize and improve fairness in medical VLM benchmarking.

Abstract

We introduce MultiMedEval, an open-source toolkit for fair and reproducible evaluation of large, medical vision-language models (VLM). MultiMedEval comprehensively assesses the models' performance on a broad array of six multi-modal tasks, conducted over 23 datasets, and spanning over 11 medical domains. The chosen tasks and performance metrics are based on their widespread adoption in the community and their diversity, ensuring a thorough evaluation of the model's overall generalizability. We open-source a Python toolkit (github.com/corentin-ryr/MultiMedEval) with a simple interface and setup process, enabling the evaluation of any VLM in just a few lines of code. Our goal is to simplify the intricate landscape of VLM evaluation, thus promoting fair and uniform benchmarking of future models.