Elie Abitbol, Alexandra Miere, Jean- Baptiste Excoffier, Carl-Joe Mehanna, Francesca Amoroso, Samuel Kerr, Matthieu Ortala, Eric H Souied

*Author details available on the publication

BMJ Open Ophtalmology
Publié le : 18 janvier 2022

What is already known about this subject?
Ultra-wide field imaging has been previously used to distinguish retinal vascular diseases from controls using deep learning, but no study has aimed at distinguishing multiple retinal vascular diseases.

What are the new findings?
By using a deep learning classifier, multiple retinal vascular diseases may be distinguished, with an accuracy of 88.4%.

How might these results change the focus of research or clinical practice?
A deep learning classifier may be a useful tool in areas with a shortage of ophthalmic care.

Other resources you may be interested in

Medical imaging: Explainability methods for image classification

Medical imaging: Explainability methods for image classification

Automatic image analysis has seen its performance grow strongly in recent years. These recent advances improve the construction of predictive imaging models, increasing their reliability

Neural Network Architectures for image classification

Neural Network Architectures for image classification

From examples where each image is associated with a category, the so-called Machine Learning models learn to identify patterns specific to the observations of the same…