Scholarlink

This study surveys the existing literature on DR detection methods, highlighting the effectiveness of various deep learning models like DiaNet, Inception, MobileNet, ResNet, Nesnet, and self-designed CNN, achieving classification accuracy ranging from 83 to 90%. The widely used APTOS 2019 Blindness Detection dataset was employed in numerous studies, supplemented by a custom dataset developed in collaboration with Aditya Birla Hospital, comprising around 302 images at different stages with necessary preprocessing. These findings emphasize the potential of deep learning to enhance DR detection accuracy, efficiency, and consistency, thereby facilitating early intervention and preserving vision.

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

Survey on Transfer Learning for Enhanced Diabetic Retinopathy Detection

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Utkarsh Jaulkar,
  • Sneha Chavan,
  • Vaibhavi Sawant,
  • Rushikesh Karande,
  • Vinodkumar Bhutnal,
  • Dhanashri Joshi

摘要

This study surveys the existing literature on DR detection methods, highlighting the effectiveness of various deep learning models like DiaNet, Inception, MobileNet, ResNet, Nesnet, and self-designed CNN, achieving classification accuracy ranging from 83 to 90%. The widely used APTOS 2019 Blindness Detection dataset was employed in numerous studies, supplemented by a custom dataset developed in collaboration with Aditya Birla Hospital, comprising around 302 images at different stages with necessary preprocessing. These findings emphasize the potential of deep learning to enhance DR detection accuracy, efficiency, and consistency, thereby facilitating early intervention and preserving vision.