Development and validation of a deep learning image quality feedback system for infant fundus photography

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...

Full description

Saved in:
Bibliographic Details
Main Authors: Helei Wang, Longhui Li, Wenjuan Wang, Zhiwen Li, Tianzi Jian, Xueying Yang, Boxuan Song, Shiqiang Li, Fabao Xu, Shaopeng Liu, Ying Li
Format: Article
Language:English
Published: Nature Portfolio 2025-07-01
Series:Scientific Reports
Subjects:
Online Access:https://doi.org/10.1038/s41598-025-10859-5
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849763796264943616
author Helei Wang
Longhui Li
Wenjuan Wang
Zhiwen Li
Tianzi Jian
Xueying Yang
Boxuan Song
Shiqiang Li
Fabao Xu
Shaopeng Liu
Ying Li
author_facet Helei Wang
Longhui Li
Wenjuan Wang
Zhiwen Li
Tianzi Jian
Xueying Yang
Boxuan Song
Shiqiang Li
Fabao Xu
Shaopeng Liu
Ying Li
author_sort Helei Wang
collection DOAJ
description 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.
format Article
id doaj-art-31b9b92ff45541a79e928fed94199bfa
institution DOAJ
issn 2045-2322
language English
publishDate 2025-07-01
publisher Nature Portfolio
record_format Article
series Scientific Reports
spelling doaj-art-31b9b92ff45541a79e928fed94199bfa2025-08-20T03:05:18ZengNature PortfolioScientific Reports2045-23222025-07-0115111210.1038/s41598-025-10859-5Development and validation of a deep learning image quality feedback system for infant fundus photographyHelei Wang0Longhui Li1Wenjuan Wang2Zhiwen Li3Tianzi Jian4Xueying Yang5Boxuan Song6Shiqiang Li7Fabao Xu8Shaopeng Liu9Ying Li10School of Instrumentation and Optoelectronic Engineering, Beihang UniversityState Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen UniversityDepartment of Ophthalmology, Qilu Hospital of Shandong UniversityDepartment of Ophthalmology, Qilu Hospital of Shandong UniversityDepartment of Hematology, Qilu Hospital, Shandong UniversityDepartment of Ophthalmology, Qilu Hospital of Shandong UniversityDepartment of Ophthalmology, Qilu Hospital of Shandong UniversityDepartment of Ophthalmology, Qilu Hospital of Shandong UniversityDepartment of Ophthalmology, Qilu Hospital of Shandong UniversitySchool of Computer Science, Guangdong Polytechnic Normal UniversityDepartment of Dermatology, Qilu Hospital of Shandong UniversityAbstract 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.https://doi.org/10.1038/s41598-025-10859-5Infant fundus imageImage qualityCapturing issuesDeep learningSample balanced strategy
spellingShingle Helei Wang
Longhui Li
Wenjuan Wang
Zhiwen Li
Tianzi Jian
Xueying Yang
Boxuan Song
Shiqiang Li
Fabao Xu
Shaopeng Liu
Ying Li
Development and validation of a deep learning image quality feedback system for infant fundus photography
Scientific Reports
Infant fundus image
Image quality
Capturing issues
Deep learning
Sample balanced strategy
title Development and validation of a deep learning image quality feedback system for infant fundus photography
title_full Development and validation of a deep learning image quality feedback system for infant fundus photography
title_fullStr Development and validation of a deep learning image quality feedback system for infant fundus photography
title_full_unstemmed Development and validation of a deep learning image quality feedback system for infant fundus photography
title_short Development and validation of a deep learning image quality feedback system for infant fundus photography
title_sort development and validation of a deep learning image quality feedback system for infant fundus photography
topic Infant fundus image
Image quality
Capturing issues
Deep learning
Sample balanced strategy
url https://doi.org/10.1038/s41598-025-10859-5
work_keys_str_mv AT heleiwang developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT longhuili developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT wenjuanwang developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT zhiwenli developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT tianzijian developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT xueyingyang developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT boxuansong developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT shiqiangli developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT fabaoxu developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT shaopengliu developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography
AT yingli developmentandvalidationofadeeplearningimagequalityfeedbacksystemforinfantfundusphotography