The early visual screening of children plays a critical role in promoting visual development, especially for those with visual impairments. Among various approaches, eye-tracking based visual assessment has emerged as a promising tool, particularly for infants, toddlers, and children with developmental disabilities who are unable to complete traditional vision tests. The object of this study is to design and investigate the effectiveness of using a deep learning based, individualized eye-tracking system to assess visual function, specifically visual acuity and visual field, in typical preschool children and infants under the age of three. This study aims to establish a reliable, noninvasive visual screening method that accommodates the diverse needs and abilities of young children.
The investigators will recruit a total of 1,300 participants across four key groups to support the development and validation of an individualized visual function screening system using eye-tracking technology: 1. 300 adults to establish a reference dataset for gaze estimation models 2. 300 typically developing preschool children aged 3 to 5 years 3. 400 typically developing children under the age of 3 4. 300 children with special needs, including those with developmental disabilities and visual impairments Data collection will take place over 15 to 22 months, depending on the group.
Study Type
OBSERVATIONAL
Enrollment
1,300
no interventions
Department of Ophthalmology, National Taiwan University Hospital
Taipei, Taiwan
RECRUITINGAccuracy of Visual Acuity Classification by Deep Vision Tracking System (DVR System), Compared to Standard Clinical Assessment
Compare the DVR system's classification (normal vs. impaired vision) with results from standard clinical tests (e.g., Teller Acuity Cards). Report sensitivity and specificity. Cut-off values: logMAR 0.22 (Snellen decimal: 0.6) and 0.097 (Snellen decimal: 0.8).
Time frame: 2 months
Visual Field Detection Accuracy of the DVR System Compared to Standard Evaluation
This outcome assesses the accuracy of the DVR system in detecting visual field response in children, defined by whether the participant successfully fixates on screen-displayed peripheral stimuli. The DVR system's outputs will be compared to clinical observation. Accuracy will be evaluated via sensitivity and specificity.
Time frame: 2 months
Criterion validity of DVR system visual acuity estimates (logMAR)
Compare DVR system logMAR estimates with clinical measurements (e.g., Teller Acuity Cards) using linear regression. Report slope coefficient; a slope near 1 indicates strong agreement. Target criterion: slope ≥ 0.8.
Time frame: 2 months
Test-retest reliability of DVR system visual acuity measurements
Participants will be re-tested within 2 weeks. Reliability will be assessed using standard error of measurement and coefficient of variation from panel data.
Time frame: 2 months
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.