CAMEL: Lightweight Retinal Foundation Models via Masked Embedding Learning
November 2, 2026
CAMEL introduces cross-architecture masked embedding learning to create ultra-lightweight retinal foundation models. The method aims to bridge the gap between high-performance ophthalmologic screening and the hardware constraints of real-world clinical environments.
HOW THIS AFFECTS YOU
●
researcherYou can leverage this cross-architecture approach to achieve SOTA performance on edge-constrained medical hardware.
●
healthThis enables more scalable deployment of advanced diagnostic AI in clinics with limited computational resources.