Can a Natural Image-Based Foundation Model Outperform a Retina-Specific Model in Detecting Ocular and Systemic Diseases?

Purpose: DINOv2 is a natural image-based foundation model (FM), pretrained exclusively on 142 million natural images from the LVD-142M data set. In contrast, RETFound is a retina-specific FM, pretrained on ∼3 million images, including natural images, color fundus photos, and OCT images (∼1 million e...

Fuld beskrivelse

Bibliografiske detaljer

Principais autores:	Qingshan Hou, PhD, Yukun Zhou, PhD, Jocelyn Hui Lin Goh, BEng, Ke Zou, PhD, Samantha Min Er Yew, BSc, Sahana Srinivasan, BEng, Meng Wang, PhD, Thaddaeus Wai Soon Lo, BEng, Xiaofeng Lei, MSc, Siegfried K. Wagner, MD, PhD, Mark A. Chia, MD, PhD, Gabriel Dawei Yang, MD, PhD, Hongyang Jiang, PhD, An Ran Ran, MD, PhD, Rui Santos, PhD, Gabor Mark Somfai, MD, Juan Helen Zhou, PhD, Haoyu Chen, MD, Qingyu Chen, PhD, Carol Y. Cheung, PhD, Pearse A. Keane, MD, Yih Chung Tham, PhD
Format:	Artigo
Sprog:	Inglês
Udgivet:	Elsevier 2026-01-01
Serier:	Ophthalmology Science
Fag:	Artificial intelligence Foundation models Ocular diseases Oculomics Retina
Online adgang:	http://www.sciencedirect.com/science/article/pii/S2666914525002210
Tags:	Tilføj Tag Ingen Tags, Vær først til at tagge denne postø!