Glaucoma Detection and Classification Using Improved U-Net Deep Learning Model.

Area of Science:

• Ophthalmology
• Medical Imaging
• Artificial Intelligence

Background:

• Glaucoma affects millions globally, with significant misdiagnosis rates in developing healthcare systems.
• Early detection of glaucoma is crucial for preventing vision loss and managing the condition effectively.
• Current diagnostic methods can be limited, necessitating advanced technological solutions.

Purpose of the Study:

• To develop and evaluate a deep learning system for the early detection and prediction of glaucoma.
• To utilize retinal fundus images for identifying glaucoma and assessing its presence in patients.
• To improve the accuracy and efficiency of glaucoma diagnosis through artificial intelligence.

Main Methods:

• A deep learning algorithm integrating pretrained transfer learning models with the U-Net architecture was employed.
• DenseNet-201, a deep convolutional neural network (DCNN), was used for feature extraction from retinal images.
• The DCNN approach was applied to segment the optic cup and classify glaucoma presence.

Main Results:

• The proposed deep learning model achieved high accuracy in detecting glaucoma from retinal fundus images.
• The model demonstrated a training accuracy of 98.82% and a testing accuracy of 96.90%.
• Comparative analysis showed the proposed model outperformed existing deep learning-based classification methods.

Conclusions:

• The developed deep learning system shows significant potential for accurate and early glaucoma detection.
• This AI-driven approach can aid in timely diagnosis, especially in regions with limited healthcare infrastructure.
• The findings suggest a more successful and reliable paradigm for glaucoma screening and diagnosis.