Glaucoma Classification Techniques Using Optic Disc Imaging

Glaucoma is a severe ocular condition that can lead to irreversible blindness if not diagnosed early. Recent advances in medical imaging and machine learning have paved the way for automated glaucoma classification systems that leverage optic disc imaging to detect this disease with high accuracy. In this study, we propose a novel classification technique using convolutional neural networks (CNNs) that integrates spatial and spectral features extracted from optic disc images. Our methodology involves pre-processing steps such as image normalization and augmentation to enhance the feature extraction process. The CNN model is trained on a dataset comprising 5,000 optic disc images, achieving a classification accuracy of 94.5%, sensitivity of 92%, and specificity of 96%. These results signify a substantial improvement over traditional image processing techniques, which typically achieve accuracies around 87%. The proposed system not only enhances glaucoma detection but also contributes to the scalability of screening processes, facilitating early intervention and treatment planning.