MedSapiens: Repurposing Human Pose Tech for Smarter Medical Landmarks
What if human pose tech could guide doctors in scans?
MedSapiens takes a simple but bold idea: adapt a powerful human pose foundation model to find anatomical landmarks—key points—in medical images.
- Built by fine-tuning Sapiens across multiple medical datasets.
- Sets a new state of the art in success detection rate (SDR): up to 5.26% better than generalist models and up to 21.81% better than specialist models.
- Works even with few labels, improving few-shot SDR by 2.69%.
Why it matters: Landmark detection underpins many clinical and research workflows—from aligning scans to measuring structures. By reusing a pose model tuned for precise spatial localization, MedSapiens shows that strong priors can outperform bespoke, task-specific designs.
Paper: http://arxiv.org/abs/2511.04255v1
Code & weights: https://github.com/xmed-lab/MedSapiens
Authors: Marawan Elbatel, Anbang Wang, Keyuan Liu, Kaouther Mouheb, Enrique Almar-Munoz, Lizhuo Lin, Yanqi Yang, Karim Lekadir, Xiaomeng Li
Paper: http://arxiv.org/abs/2511.04255v1
Register: https://www.AiFeta.com
AI MedicalImaging ComputerVision PoseEstimation FoundationModels OpenSource Research Healthcare
