Overcoming Data Scarcity in Biomedical Imaging with a Foundational Multi-Task Model

TL;DR

This paper introduces UMedPT, a multi-task foundational model trained on diverse biomedical imaging datasets, which achieves high performance with minimal data and enhances cross-center transferability.

Contribution

The authors develop UMedPT, a multi-task foundational model for biomedical imaging that reduces data requirements and improves transferability across different imaging domains.

Findings

UMedPT outperforms ImageNet pretraining and previous models.

It maintains performance with only 1% of training data on in-domain tasks.

It requires no more than 50% of training data for out-of-domain tasks.

Abstract

Foundational models, pretrained on a large scale, have demonstrated substantial success across non-medical domains. However, training these models typically requires large, comprehensive datasets, which contrasts with the smaller and more heterogeneous datasets common in biomedical imaging. Here, we propose a multi-task learning strategy that decouples the number of training tasks from memory requirements. We trained a Universal bioMedical PreTrained model (UMedPT) on a multi-task database including tomographic, microscopic, and X-ray images, with various labelling strategies such as classification, segmentation, and object detection. The UMedPT foundational model outperformed ImageNet pretraining and the previous state-of-the-art models. For tasks related to the pretraining database, it maintained its performance with only 1% of the original training data and without fine-tuning. For out-of-domain tasks it required not more than 50% of the original training data. In an external independent validation imaging features extracted using UMedPT proved to be a new standard for cross-center transferability.