Scholarlink

Diabetic eye diseases (DED), which comprise Cataract, Glaucoma, Diabetic Macular Edema (DME), and Diabetic Retinopathy (DR), are posing a great threat to the sight condition in case of late diagnosis. This work exploits the cutting-edge deep learning models that are at the forefront of modern advancements which are VGG16, EfficientNet-B0, ResNet152V2, and Bi-GRU + ResNet152V2 hybrid-for classification of the mentioned DED using retinal fundus images. Among them, precision and recall values of 95% ,with an accuracy result of 95%. The techniques of transfer learning and data augmentation were used to boost up the performance of the model. Therefore, the proposed systems are strongly feasible with regard to the early detection of DED and also provide a robust instrument assisting the ophthalmologists at diagnosis. This work therefore contributes toward growing research toward enhancing automated medical diagnostics using deep learning.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

DeepDiabetic: Detecting Diabetic Eye Diseases with Neural Networks

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Nukala Vijay Kumar,
  • Sireesha Moturi,
  • Tadala Manideep,
  • Jupalli Rajesh,
  • Manikala Manikanta Rambabu,
  • Mothe Sathyam Reddy

摘要

Diabetic eye diseases (DED), which comprise Cataract, Glaucoma, Diabetic Macular Edema (DME), and Diabetic Retinopathy (DR), are posing a great threat to the sight condition in case of late diagnosis. This work exploits the cutting-edge deep learning models that are at the forefront of modern advancements which are VGG16, EfficientNet-B0, ResNet152V2, and Bi-GRU + ResNet152V2 hybrid-for classification of the mentioned DED using retinal fundus images. Among them, precision and recall values of 95% ,with an accuracy result of 95%. The techniques of transfer learning and data augmentation were used to boost up the performance of the model. Therefore, the proposed systems are strongly feasible with regard to the early detection of DED and also provide a robust instrument assisting the ophthalmologists at diagnosis. This work therefore contributes toward growing research toward enhancing automated medical diagnostics using deep learning.