Scholarlink

Diabetic Retinopathy (DR) is a leading cause of preventable blindness, affecting over 100 million adults globally, with prevalence expected to rise significantly by 2045. Early detection of DR and diabetic macular edema (DME) is crucial for timely treatment . This study addresses the challenges of DR diagnosis using ultra-widefield (UWF) fundus images, which provide a 200-degree retinal view. We participated in the MICCAI 2024 UWF4DR challenge, tackling three tasks: image quality assessment, DR detection, and DME identification, using AutoMorph, ShuffleNet, and EfficientNetB0 models, respectively. Our approach secured competitive leaderboard rankings, highlighting the potential of deep learning models for scalable, automated DR diagnosis in clinical settings.

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

Lightweight and Accurate: ShuffleNet for Diabetic Retinopathy and EfficientNet for Diabetic Macular Edema Diagnosis

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Berthold Scheuringer,
  • Moritz Haderer,
  • Martin Marinschek,
  • Oleksandra Menzatiuk,
  • Vera Pils

摘要

Diabetic Retinopathy (DR) is a leading cause of preventable blindness, affecting over 100 million adults globally, with prevalence expected to rise significantly by 2045. Early detection of DR and diabetic macular edema (DME) is crucial for timely treatment . This study addresses the challenges of DR diagnosis using ultra-widefield (UWF) fundus images, which provide a 200-degree retinal view. We participated in the MICCAI 2024 UWF4DR challenge, tackling three tasks: image quality assessment, DR detection, and DME identification, using AutoMorph, ShuffleNet, and EfficientNetB0 models, respectively. Our approach secured competitive leaderboard rankings, highlighting the potential of deep learning models for scalable, automated DR diagnosis in clinical settings.