EndoDINO, a DINOv2-style model pre-trained on up to 10 million curated endoscopy images, reports strong or state-of-the-art results on three GI endoscopy benchmarks with frozen features and simple heads.
Acosta, Zahra Shakeri, John P.A
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
EndoDINO: A Foundation Model for GI Endoscopy
EndoDINO, a DINOv2-style model pre-trained on up to 10 million curated endoscopy images, reports strong or state-of-the-art results on three GI endoscopy benchmarks with frozen features and simple heads.