HOPPR launches mammography vision-language model

HOPPR has introduced the HOPPR EB 2D Mammo Narrative Model, a vision-language model designed to translate 2D mammography images into narrative language describing imaging characteristics.

Designed for developers, the model is positioned as a foundational software component for AI-assisted breast imaging and radiology workflow applications.

Structured JSON output is generated from standard 2D mammography images. HOPPR said the output can be integrated into downstream radiology workflow applications.

Training included more than 200,000 mammography studies from multiple U.S. sites. The dataset covered varied breast density categories and implant-displaced imaging scenarios, according to the company.

Version control allows developers to lock specific model versions for consistency during application development and updates. HOPPR also said training data records are maintained to support traceability and bias assessment across the model lifecycle.

Access is available through HOPPR’s Forward Deployed Services team, which supports partners in evaluating the model and configuring integrations for specific workflows and data environments.

The release expands HOPPR’s vision-language model portfolio and follows its recent chest radiography narrative model.