University of Michigan VLM trained on ~1M clinical neuroimaging studies outperforms general and medical models
mlinslab · hf · 2026-11-14
A University of Michigan team (ML in Neurosurgery Lab and collaborators) presents a frontier vision-language model for human neuroimaging, trained on nearly one million clinical neuroimaging studies. It interprets complete examinations and discusses findings in natural language, reportedly outperforming existing general and medical models on diagnosis, lesion localization, reporting, and progression assessment. Project page: neurovl.mlins.org.
More from Apps
- Claude Meetings sidelines itself in client calls, treating every customer as an untrusted outsider — dotey · 2026-11-14
- Google Cloud launches Gemini for healthcare to unify clinical data and automate workflows — OnlineInference · 2026-11-14
- 1M spoken words later: voice dictation now beats typing for this user — danshipper · 2026-11-14
- Haru: open-source realtime voice chat app inspired by the movie HER — cocktailpeanut · 2026-11-14
- Run a private AI fully offline on your phone with PocketPal AI in 5 steps — alex_verem · 2026-11-14
- Rand Fishkin hammers the point: AI citations are not recommendations — sanderssays · 2026-11-14