This study explores the use of deep learning (DL) for automated detection of Type 1 retinopathy of prematurity (ROP), Type 2 ROP, and aggressive ROP (A-ROP) using fundus images. A RegNetY002 convolutional neural network (CNN) model was trained on 634 fundus images from 317 premature infants. The model demonstrated high accuracy and specificity in distinguishing between different ROP types and stages, with AUC scores ranging from 0.85 to 0.98. The integration of this DL-based AI algorithm into clinical practice could potentially reduce ophthalmologist workload and improve decision-making in ROP management.
Publisher
Eye
Published On
Jun 25, 2024
Authors
Eşay Kıran Yenice, Caner Kara, Çağatay Berke Erdaş
Tags
deep learning
retinopathy of prematurity
fundus images
convolutional neural network
automated detection
neonatal care
medical imaging
Related Publications
Explore these studies to deepen your understanding of the subject.
