Michael J. Pazzani

Comparison of Deep Learning and Clinician Performance for Detecting Referable Glaucoma from Fundus Photographs in a Safety Net Population.

A deep learning algorithm effectively detects referable glaucoma from fundus photographs, outperforming clinicians with variable diagnostic accuracy. This tool could improve screening efficiency and access to timely glaucoma care.

Detecting Glaucoma from Fundus Photographs Using Deep Learning without Convolutions: Transformer for Improved Generalization.

Vision Transformers improve glaucoma detection from fundus photos, outperforming ResNet-50 in generalizability across diverse datasets. This enhances diagnostic accuracy and explainability in ophthalmic imaging.