Glaucoma Classification from Fundus Images Using Meta-Learning Models

The present paper introduces a novel method for the accurate classification of glaucoma with minimal labeled samples using a combination of meta-learning and data augmentation techniques. The proposed model is trained to learn generalized features from fundus images and can efficiently adapt to new, unseen classes with only a few labeled examples. Extensive experiments on the 1000 Fundus image dataset demonstrate the improved classification accuracy and robustness of the proposed method over existing few-shot learning techniques and traditional deep learning models. The results underline the potential of few-shot learning for real-world applications in glaucoma screening, particularly in environments where availability of large, labeled datasets is restricted.