OmniRad: A Radiological Foundation Model for Multi-Task Medical Image Analysis

TL;DR

OmniRad is a large self-supervised radiological foundation model trained on 1.2 million images, demonstrating strong transferability across various medical imaging tasks and modalities.

Contribution

The paper introduces OmniRad, a radiology-inspired foundation model that emphasizes representation reuse and cross-task transferability for medical image analysis.

Findings

Improves classification F1 by up to 2.05% on MedMNISTv2.

Achieves higher Dice scores on MedSegBench datasets with frozen representations.

Shows enhanced feature clustering and modality separation in visualizations.

Abstract

Radiological analysis increasingly benefits from pretrained visual representations that can support heterogeneous downstream tasks across imaging modalities. In this work, we introduce OmniRad, a self-supervised radiological foundation model pretrained on 1.2 million medical images, designed with radiology-inspired principles emphasizing representation reuse and cross-task transferability. We evaluate the pretrained encoder under multiple downstream adaptation regimes, including lightweight task-specific adapters with a frozen backbone as well as full end-to-end fine-tuning for classification, allowing us to assess both representation quality and task-specific performance. OmniRad is evaluated on a broad suite of public benchmarks spanning classification and segmentation across multiple modalities. On the MedMNISTv2 collection, OmniRad improves classification F1 by up to 2.05% over competing foundation models. For dense prediction, OmniRad attains mean Dice score improvements across six MedSegBench datasets when using frozen representations. Qualitative analyses and latent-space visualizations suggest improved feature clustering and modality-related separation.