3.8 Proceedings Paper

ELEVATING FUNDOSCOPIC EVALUATION TO EXPERT LEVEL - AUTOMATIC GLAUCOMA DETECTION USING DATA FROM THE AIROGS CHALLENGE

Publisher

IEEE
DOI: 10.1109/ISBIC56247.2022.9854758

Keywords

ophthalmology; glaucoma detection; airogs challenge

Ask authors/readers for more resources

Glaucoma is a eye disease that leads to vision loss and is usually asymptomatic until its late stages. In this study, an artificial intelligence approach based on ocular fundus images was used for glaucoma detection. By combining object detection and classification networks, a high sensitivity for glaucoma detection was achieved.
Glaucoma is a group of clinically relevant eye diseases that eventually result in damage to the retina and optic nerve, leading to vision loss. While it is the main cause of irreversible blindness worldwide, it typically remains asymptomatic until its late stages. Glaucoma is usually diagnosed during routine eye examination, which includes fundoscopic evaluation. Here, we describe our approach for glaucoma detection in the Artificial Intelligence for Robust Glaucoma Screening challenge based on ocular fundus images. We first use object detection to focus on the most relevant part of the image and then use an ensemble of classification networks. Ungradability is rated by combining the reliability score during object detection with an explicit rating of an ungradability neural network. We achieve a sensitivity of detecting glaucoma of 0.8396 at 95% specificity and an area under the curve of 0.9852 for ungradability detection.

Authors

I am an author on this paper
Click your name to claim this paper and add it to your profile.

Reviews

Primary Rating

3.8
Not enough ratings

Secondary Ratings

Novelty
-
Significance
-
Scientific rigor
-
Rate this paper

Recommended

No Data Available
No Data Available