Automated Detection of Glaucoma and Diagnostic Features for Justraigs Challenge

Glaucoma encompasses a set of ocular conditions that jeopardize vision and can lead to blindness by harming the optic nerve at the rear of the eye. While early detection is crucial to halt further vision impairment, it is often challenging due to the scarcity of medical experts. To mitigate this lim...

Full description

Saved in:
Bibliographic Details
Main Author: Kubrak, Tomasz
Format: Conference Proceeding
Language:English
Subjects:
Biomedical optical imaging
Blindness
Deep architecture
Diagnostic Features Classification
Feature extraction
Glaucoma
Glaucoma Detection
JustRAIGS Challenge
Pipelines
Visual impairment
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Glaucoma encompasses a set of ocular conditions that jeopardize vision and can lead to blindness by harming the optic nerve at the rear of the eye. While early detection is crucial to halt further vision impairment, it is often challenging due to the scarcity of medical experts. To mitigate this limitation, various deep learning architectures have been developed. However, due to their importance, further improvement of these models is crucial. They also tend to operate as "black boxes", which makes their decision process opaque. To address these issues, our work not only establishes a new state-of-the-art (SOTA) glaucoma classification pipeline on JustRAIGS dataset but also detects ten additional features commonly used by ophthalmologists to explain their diagnostic decisions.
ISSN:1945-8452
DOI:10.1109/ISBI56570.2024.10635144