Integrated Deep Learning Framework for Automated Glaucoma Detection, Optic Disc/ Cup Segmentation, and CDR Calculation

• Published in: 2025 International Conference on Visual Analytics and Data Visualization (ICVADV) pp. 887 - 894
• Main Authors: K J, Tina Brivitha, S, Sophia
• Format: Conference Proceeding
• Language: English
• Published: IEEE 04.03.2025
• Subjects: Accuracy; Automatic optical inspection; Biomedical optical imaging; Convolutional Neural Networks (CNN); Cup to Disc Ratio (CDR); Deep learning; Glaucoma; Glaucoma Detection; Image segmentation; Integrated optics; Manuals; Medical Images; Optic Disc Segmentation; Optical fibers; Optical imaging
• DOI: 10.1109/ICVADV63329.2025.10961500

The earliest stage of ocular glaucoma is often the leading cause of irreversible vision loss globally, primarily due to late diagnosis and its typically asymptomatic nature. This makes early diagnosis and timely intervention critical for preventing or at least partially slowing vision deterioration. Traditional diagnostic techniques, which are often time-consuming, remain widely used. However, manual estimation by ophthalmologists frequently leads to significant variability in intra- and inter-subject results. This paper proposes a novel deep learning-based system for optic disc (OD) and optic cup (OC) segmentation, as well as cup-to-disc ratio (CDR) calculation, to approximate glaucoma levels. The system utilizes a CNN-based approach that ensures accurate segmentation of the nerve fiber layer, reduces reliance on manual CDR inspection, and facilitates early glaucoma diagnosis.