The Early Diagnosis of Retinal Diseases for Cataracts and Glaucoma Detection Using a Deep Learning-based Approach.
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Although it might be difficult and error-prone, identifying eye problems is essential. Modern equipment is required for doctors to promptly and precisely diagnose eye disorders. The majority of earlier research has only used deep learning models to detect or classify datasets related to eye diseases. The lack of the evaluation performance of a single end-to-end task and the detection plus classification is found. The study uses the YOLOv8 end-to-end detection model and the hybrid YOLOv8+CNN pipeline, which detects the ROI and crops the image to enter the CNN model to identify eye diseases. A dataset of retinal eye diseases, including cataracts, glaucoma, and healthy eyes, is used in this work. According to this study, the YOLOv8 model achieved a mAP50 of 92% and surpassed the hybrid YOLOv8+CNN model by 89%. This suggests that sensitivity of error propagation, ROI-based detection, and dataset size when using the hybrid models.
Fields of Study & Subjects
Retinal Imaging and Analysis
Advanced Neural Network Applications
Ophthalmology and Visual Health Research
Publication details
- Journal / source
- Zanin Journal of Science and Engineering
- Volume
- 2
- Issue
- 1
- Pages
- 01–11
- Publication date
- 30 March 2026
- Language
- EN