Related Experiment Video
Updated: Oct 31, 2025
05:36
Subjective Refraction Test Using a Smartphone for Vision Screening
Published on: October 18, 2024
1.3K
Machine Learning-Guided Prediction of Central Anterior Chamber Depth Using Slit Lamp Images from a Portable
David Chen1, Yvonne Ho2, Yuki Sasa2
1Department of Ophthalmology, National University Hospital, Singapore 119228, Singapore.
Biosensors
|July 2, 2021
Summary
This study shows that machine learning can estimate central anterior chamber depth (ACD) using smartphone slit lamp images. This offers a potential portable screening tool for narrow angles.
Area of Science:
- Ophthalmology
- Medical Imaging
- Artificial Intelligence
Background:
- Narrow angles pose a risk for angle-closure glaucoma.
- Current screening methods for narrow angles lack portability and objectivity in community settings.
Purpose of the Study:
- To evaluate the feasibility of using machine learning on portable slit lamp images to predict central anterior chamber depth (ACD).
- To develop a non-invasive, community-based screening method for narrow angles.
Main Methods:
- A prospective, single-centre image validation study.
- Slit lamp images were captured using a portable smartphone device (MIDAS) in phakic patients (≥60 years, no prior eye surgery).
- Machine learning algorithms processed images to predict ACD, comparing results with anterior segment optical coherence tomography (ASOCT).
Main Results:
- A strong positive correlation was observed between predicted ACD and ASOCT-measured ACD (R² = 0.91 for training, R² = 0.73 for test data).
- The study successfully validated the use of machine learning for ACD estimation from portable device images.
Conclusions:
- Portable smartphone-based slit lamp imaging combined with machine learning shows promise for estimating central ACD.
- This technology could enable objective, portable screening for narrow angles in community settings, aiding early glaucoma detection.
