Advanced Glaucoma Detection by Integrating ANFIS with CNN Models for Enhanced Diagnosis

Glaucoma is a retinal disease affecting millions of people round the globe. It can lead to irreversible blindness unless detected early. In this research we differentiate glaucomatous eyes from normal eyes using a novel glaucoma detection framework that employs Adaptive Neuro-Fuzzy Inference System classifier, VGG-19 and ResNet50 models. The proposed method concatenates the classification results from ANFIS, VGG-19 and ResNet50 models using post processing decision block that provides the final classified output. This approach overcomes the limitations encountered in existing techniques which significantly enhance diagnostic accuracy and robustness by incorporating the merits of the aforesaid three models. We precisely develop the framework through ensemble modelling which results an accuracy of 95.40%, a precision rate of 99.40%, a recall rate of 88.30% and an F1-score of 93.50%. Our research findings highlight the significance of integrating conventional optic disc measurements methods with image shape-derived features in machine learning models. This enables the early and accurate glaucoma diagnosis.