Deep Learning Model for Binary Glaucoma Diagnosis from Eye Fundus Images

Glaucoma is one of the leading causes of blindness worldwide, characterized by progressive degeneration of the optic nerve and often associated with elevated intraocular pressure. Early detection is crucial to prevent irreversible vision loss, yet traditional diagnostic methods require specialized equipment and expertise, limiting accessibility. This work presents deep learning models implementations for the automated binary classification of glaucoma using eye fundus images. The study explores experimental set up and data set preprocessing to optimize model performance. Results demonstrate that a balanced approach to augmentation and fine-tuning significantly enhances the model's accuracy and generalizability. The findings suggest that this model can serve as an effective toool to support diagnosis, particularly in regions with limited access to ophthalmological specialists. The source code for replicating the experiments is publicly available at https://github.com/valentinaache/glaucomaPGUSab.