Deep Learning for Diabetic Retinopathy Detection

This presentation addresses diabetic retinopathy (DR) detection using deep learning, highlighting its importance for early diagnosis and improved patient outcomes. It introduces DR challenges, the APTOS 2019 dataset, and evaluates models such as ResNet18, MobileNetV2, and custom CNNs for multi-class severity classification and binary screening. Methodologies, performance metrics, and trade-offs between complexity and accuracy are discussed. The analysis examines limitations like dataset...