Scholarlink

Diabetes is a major worldwide health concern since, by 2030, the World Health Organization predicts it will rank seventh in terms of cause of death. The goal of this research is to use machine learning and deep convolutional neural networks to create an automated detection system. In this research, deep learning and transfer learning algorithms are used to analyze the stages of diabetic retinopathy. Transfer learning extracts features from trained models, using techniques like CNNs and deep learning for DR image classification, particularly for patient DR images, enhancing performance. Using a dataset of 3662 train images, it employs CNN, hybrid CNN with ResNet, and hybrid CNN with DenseNet. The models’ respective accuracy percentages are 96.22%, 93.18%, and 75.61%. In the comparative analysis, hybrid CNN with DenseNet is shown to be the best deep learning classification model for automated DR detection.

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

Diabetic Retinopathy Stages Classification Using Advance Deep Learning Models

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Boina Sirisha

摘要

Diabetes is a major worldwide health concern since, by 2030, the World Health Organization predicts it will rank seventh in terms of cause of death. The goal of this research is to use machine learning and deep convolutional neural networks to create an automated detection system. In this research, deep learning and transfer learning algorithms are used to analyze the stages of diabetic retinopathy. Transfer learning extracts features from trained models, using techniques like CNNs and deep learning for DR image classification, particularly for patient DR images, enhancing performance. Using a dataset of 3662 train images, it employs CNN, hybrid CNN with ResNet, and hybrid CNN with DenseNet. The models’ respective accuracy percentages are 96.22%, 93.18%, and 75.61%. In the comparative analysis, hybrid CNN with DenseNet is shown to be the best deep learning classification model for automated DR detection.