Related Experiment Video
Updated: May 9, 2026
Subjective Refraction Test Using a Smartphone for Vision Screening
Published on: October 18, 2024
Smartphone-based lightweight AI system for real-time multiple anterior segment disease screening: development and
Yuwen Liu1, Changsheng Xu1,2, Shinan Wu1
1Fujian Provincial Key Laboratory of Ophthalmology and Visual Science, Xiamen University affiliated Xiamen Eye Center, Fujian Engineering and Research Center of Eye Regenerative Medicine, Eye Institute of Xiamen University, School of Medicine of Xiamen University, Xiamen, Fujian, China.
Background:
Anterior segment diseases are a major global cause of preventable blindness, especially in regions with limited access to specialized ophthalmic care. Diagnosis typically requires slit-lamp biomicroscopy, creating a significant access bottleneck in primary care and rural settings. Existing AI solutions often lack the efficiency and generalizability necessary for widespread mobile deployment and primarily focus on individual conditions, failing to meet the demand for scalable, multi-disease screening platforms.
Methods:
We developed the Intelligent Detection System (IDS), a smartphone-compatible AI platform for real-time, automated, multi-disease screening of anterior segment diseases. The core model, Eye-YOLO, is a novel, lightweight deep learning model optimized for standard smartphone images. IDS integrates image quality assessment and urgency classification. The model was rigorously trained on a large, heterogeneous multi-center dataset of 24,671 images (comprising 17,853 slit-lamp and 6,818 smartphone images) collected from three tertiary hospitals in China. IDS was prospectively deployed and validated via a widely accessible WeChat Mini Program for community screening.
Results:
Eye-YOLO achieved a mean average precision of 0.816 with an exceptionally compact architecture (2.77 million parameters and 7.3G FLOPs), enabling real-time inference at 131 FPS on mobile devices. In controlled testing, IDS demonstrated high diagnostic performance (93.50% accuracy; AUC = 0.9837) with consistent performance across multiple smartphone brands (AUC > 0.95). Crucially, with AI assistance, junior ophthalmologists achieved an accuracy of 93.79% (AUC 0.8984), compared with 95.79% for senior experts. In real-world external validation via the WeChat Mini Program, IDS achieved 98.25% accuracy.
Conclusions:
IDS provides a highly efficient, robust, and globally scalable framework for multi-disease anterior segment screening using ubiquitous smartphone technology. Its deployment through a widely accessible mobile platform offers a promising public health solution that facilitates early detection and triage, with the potential to improve access to ophthalmic care between specialized centers and primary care settings.
Trial Registration:
Chinese Clinical Trial Registry: ChiCTR2200060808.
