Unsupervised diffusion autoencoder restores artifacts in handheld fundus images using only high-quality table-top training data and raises diagnostic accuracy to 81.17% on unseen test cases.
Review of smartphone funduscopy for diabetic retinopathy screening
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Diffusion Autoencoder for Unsupervised Artifact Restoration in Handheld Fundus Images
Unsupervised diffusion autoencoder restores artifacts in handheld fundus images using only high-quality table-top training data and raises diagnostic accuracy to 81.17% on unseen test cases.