Related Experiment Video
Updated: Sep 14, 2025
Author Spotlight: An Automated Method for Assessing Visual Acuity in Infants and Toddlers Using an Eye-Tracking System
Published on: March 17, 2023
Development and validation of a deep learning image quality feedback system for infant fundus photography
Helei Wang1,2,3,4,5, Longhui Li6, Wenjuan Wang2
1School of Instrumentation and Optoelectronic Engineering, Beihang University, Beijing, 100191, China.
Insights
A new AI system, the Deep Learning Infant Fundus Quality Feedback System (DLIF-QFS), assesses infant retinal image quality for retinopathy of prematurity (ROP) screening. This tool aids in early detection and diagnosis, addressing ophthalmologist shortages.
Area of Science:
- Ophthalmology
- Artificial Intelligence
- Medical Imaging
Background:
- Retinopathy of prematurity (ROP) is a leading cause of childhood blindness.
- Shortages of trained ophthalmologists hinder effective ROP screening.
- Automated quality assessment of infant retinal images is needed to support diagnosis.
Purpose of the Study:
- To develop and validate an AI system, the Deep Learning Infant Fundus Quality Feedback System (DLIF-QFS), for assessing infant retinal photograph quality.
- To detect operational errors in images used for ROP screening.
- To support and improve the efficiency of ROP screening and diagnosis.
Main Methods:
- Development of the DLIF-QFS using a dataset of 13,372 infant retinal images.
- Rigorous validation of the system's performance in classifying image quality (poor, adequate, excellent).
- Evaluation of the system's ability to identify image quality issues and assist in expert diagnostic tasks.
Main Results:
- The DLIF-QFS achieved high performance in overall quality classification, with AUC values of 0.802 (poor), 0.691 (adequate), and 0.926 (excellent) on an external validation dataset.
- AUC values exceeded 0.8 for most tasks identifying issues in adequate and poor quality images.
- The system demonstrated improved accuracy and consistency in expert diagnostic tests.
Conclusions:
- The DLIF-QFS is a valuable tool for assessing infant retinal image quality and identifying operational errors in ROP screening.
- The system enhances diagnostic efficiency by identifying causes of poor image quality and suggesting improvements.
- DLIF-QFS has the potential to significantly advance ROP diagnosis and management, especially in resource-limited settings.
Abstract:
Retinopathy of prematurity (ROP) is a significant cause of childhood blindness. Many healthcare institutions face a shortage of well-trained ophthalmologists for conducting screenings. Hence, we have developed the Deep Learning Infant Fundus Quality Feedback System (DLIF-QFS) to assess the overall quality of infant retinal photographs and detect common operational errors to support ROP screening and diagnosis. Our DLIF-QFS has been developed and rigorously validated using datasets comprising 13,372 images. In terms of overall quality classification, the DLIF-QFS demonstrated remarkable performance. The area under the curve (AUC) values for discriminating poor quality, adequate quality, and excellent quality images in the external validation dataset were 0.802, 0.691, and 0.926, respectively. For most classification tasks related to identifying issues in adequate and poor quality images, the AUC values consistently exceeded 0.8. In expert diagnostic tests, the DLIF-QFS improved accuracy and enhanced consistency. Its capability to identify the causes of poor image quality, enhance image quality and assist clinicians in improving diagnostic efficiency makes it a valuable tool for advancing ROP diagnosis.
