Key points are not available for this paper at this time.
The rising incidence of eye disorders due to in-creased electronic device usage highlights the necessity for ac-curate detection of anomalies in the optic disc (OD) and optic cup (OC) in retinal fundus images, crucial for the early diag-nosis of conditions such as glaucoma and diabetic retinopathy. Addressing this pivotal challenge, a novel approach is presented utilizing a modified attention-based residual U-Net architecture. Comprehensive experimentation, incorporating diverse datasets such as Drishti-GS, REFUGE, and RIM-ONE-R3, demonstrate the model's adaptability across various scenarios. Commendable performance metrics are observed, with an Intersection-over-Union (IoU) and Dice Coefficient (DC) of 87.77% and 93.48% for the optic cup, and 95.09% and 97.48% for the optic disc, respectively. These results underscore the versatility and effectiveness of the proposed methodology in achieving superior segmentation performance. Therefore, this study represents a significant advancement in retinal fundus image segmentation, enhancing early diagnosis and contributing to more effective ocular health management.
Alam et al. (Thu,) studied this question.