2026Bowang Labopen source
EchoJEPA
EchoJEPA is a latent predictive foundation model for echocardiography, trained on 18 million echocardiograms across 300,000 patients. It aims to reduce annotation burden and improve diagnostic consistency by learning generalizable representations from large unlabeled video archives. ([arxiv.org](https://arxiv.org/abs/2602.02603?utm_source=openai))
Latent predictive modeling for echocardiography
Trained on 18 million echocardiograms
Reduces left ventricular ejection fraction estimation error by 19%
pricingOpen-source and free to use
regionUnited States