Vision Foundry: A System for Training Foundational Vision AI Models

TL;DR

Vision Foundry is a user-friendly, HIPAA-compliant platform that enables clinical researchers to train and deploy foundational vision AI models using self-supervised learning, significantly reducing technical barriers and annotation needs.

Contribution

The paper introduces Vision Foundry, a code-free platform that integrates advanced SSL strategies for easy training and deployment of vision models in medical imaging.

Findings

Models trained with Vision Foundry outperform baselines in segmentation and regression tasks.

The platform demonstrates robust zero-shot generalization across different imaging protocols.

It enables domain experts to develop clinical AI tools with minimal annotation effort.

Abstract

Self-supervised learning (SSL) leverages vast unannotated medical datasets, yet steep technical barriers limit adoption by clinical researchers. We introduce Vision Foundry, a code-free, HIPAA-compliant platform that democratizes pre-training, adaptation, and deployment of foundational vision models. The system integrates the DINO-MX framework, abstracting distributed infrastructure complexities while implementing specialized strategies like Magnification-Aware Distillation (MAD) and Parameter-Efficient Fine-Tuning (PEFT). We validate the platform across domains, including neuropathology segmentation, lung cellularity estimation, and coronary calcium scoring. Our experiments demonstrate that models trained via Vision Foundry significantly outperform generic baselines in segmentation fidelity and regression accuracy, while exhibiting robust zero-shot generalization across imaging protocols. By bridging the gap between advanced representation learning and practical application, Vision Foundry enables domain experts to develop state-of-the-art clinical AI tools with minimal annotation overhead, shifting focus from engineering optimization to clinical discovery.