Related Experiment Video
Updated: Jun 10, 2026
05:40
Using an Automated Hirschberg Test App to Evaluate Ocular Alignment
Published on: March 24, 2020
12.4K
Early detection of visual impairment in young children using a smartphone-based deep learning system
Wenben Chen1, Ruiyang Li1, Qinji Yu2
1State Key Laboratory of Ophthalmology, Zhongshan Ophthalmic Center, Sun Yat-sen University, Guangdong Provincial Key Laboratory of Ophthalmology and Vision Science, Guangdong Provincial Clinical Research Center for Ocular Diseases, Guangzhou, China.
Nature Medicine
|January 26, 2023
Summary
Early detection of visual impairment in young children is challenging. A new smartphone system, Apollo Infant Sight (AIS), analyzes gazing and facial features to identify over 16 ophthalmic disorders, improving screening accessibility.
Area of Science:
- Ophthalmology
- Pediatrics
- Mobile Health (mHealth)
Background:
- Early detection of visual impairment in young children is critical but hindered by limited cooperation with standard tests.
- Traditional screening methods struggle with real-world application, despite observable features in visually impaired children.
- Existing challenges necessitate innovative, accessible screening tools for early identification of pediatric visual disorders.
Purpose of the Study:
- To develop and validate a mobile health (mHealth) system for early detection of visual impairment in infants and young children.
- To assess the efficacy of the smartphone-based Apollo Infant Sight (AIS) system in identifying children with ophthalmic disorders.
- To evaluate the system's performance in both clinical settings and at-home use by untrained caregivers.
Main Methods:
- Development and validation of the Apollo Infant Sight (AIS) system using videos from 3,652 children (≤48 months).
- AIS records and analyzes gazing behaviors and facial features under visual stimuli to detect visual impairment.
- Prospective data collection for internal and external validation, including at-home implementation tests.
Main Results:
- AIS achieved an area under the receiver operating curve (AUC) of 0.940 in internal validation and 0.843 in external validation.
- The system demonstrated strong performance in at-home use by untrained parents/caregivers, achieving an AUC of 0.859.
- AIS successfully identified children with a wide range of 16 ophthalmic disorders, adapting to diverse testing conditions.
Conclusions:
- The smartphone-based Apollo Infant Sight (AIS) system offers a promising mHealth solution for early detection of visual impairment in young children.
- AIS has the potential to significantly improve screening accessibility and accuracy for pediatric ophthalmic disorders.
- The system empowers healthcare professionals, parents, and caregivers with a tool for identifying visual impairment across various settings.
