Related Experiment Video
Updated: Aug 26, 2025
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
2.9K
Peripapillary atrophy classification using CNN deep learning for glaucoma screening
Abdullah Almansour1,2, Mohammed Alawad2,3,4, Abdulrhman Aljouie2,3,4
1Department of Imaging Research, King Abdullah International Medical Research Center, Riyadh, Saudi Arabia.
Plos One
|October 6, 2022
Summary
A new deep learning model effectively detects peripapillary atrophy (PPA), a glaucoma indicator, in fundus images. This AI tool aids ophthalmologists in early glaucoma screening and diagnosis.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Glaucoma is a leading cause of global blindness.
- Peripapillary atrophy (PPA) is a key morphological sign linked to glaucoma.
- Accurate clinical detection of PPA is crucial for timely glaucoma diagnosis.
Purpose of the Study:
- To develop a deep learning-based method for detecting PPA in fundus images.
- To create a screening tool for ophthalmologists and optometrists.
- To evaluate the model's performance using diverse datasets.
Main Methods:
- Utilized a Mask R-CNN for region of interest (ROI) localization.
- Employed a Convolutional Neural Network (CNN) for PPA classification.
- Trained and tested the model on 2,472 images from public and Saudi-based sources.
Main Results:
- Achieved Area Under the Curve (AUC) scores of 0.83 (local), 0.89 (public), and 0.87 (combined datasets).
- Demonstrated the model's efficacy in identifying PPA across different data sources.
- Indicated strong performance for the classification network.
Conclusions:
- The developed deep learning model shows promise for assisting glaucoma diagnosis in screening programs.
- Further research is recommended for PPA boundary segmentation for detailed analysis.
- Integration with optic disc and cup measurements could enhance glaucoma assessment.
