Segmentation and Classification of Glaucoma Using U-Net with Deep Learning Model.

Area of Science:

• Ophthalmology
• Computer Science
• Artificial Intelligence

Background:

• Glaucoma is a leading cause of irreversible blindness globally, affecting millions and often undetected in developing nations.
• Early diagnosis is critical to prevent vision loss, necessitating advanced detection systems.
• Current diagnostic methods can be invasive or require specialized expertise, highlighting the need for automated solutions.

Purpose of the Study:

• To develop and evaluate a deep learning model for the early prediction of glaucoma from retinal fundus images.
• To implement optic cup segmentation and feature extraction for accurate glaucoma classification.
• To assess the model's performance against established deep learning techniques.

Main Methods:

• Utilized the ORIGA dataset for training and testing the deep learning model.
• Employed U-Net architecture for optic cup segmentation and DenseNet-201 for feature extraction.
• Implemented a deep convolutional neural network (DCNN) for glaucoma classification, comparing it with VGG-19, Inception ResNet, ResNet 152v2, and DenseNet-169.

Main Results:

• The proposed deep learning model achieved high performance metrics, including 98.82% accuracy during training and 96.90% during testing.
• The model demonstrated superior performance compared to other contemporary deep learning models evaluated.
• The system effectively classifies retinal images, indicating the presence or absence of glaucoma.

Conclusions:

• The developed deep learning model shows significant promise for accurate and early glaucoma detection using retinal fundus images.
• This automated system can aid clinicians in diagnosing glaucoma, particularly in regions with limited access to specialized eye care.
• Further validation and implementation of this model could contribute to reducing the global burden of glaucoma-induced blindness.