Indicative only - not a medical device.
AcuityCheck is a web app that turns your webcam and screen into an interactive visual acuity tester. It uses computer vision to estimate your eye–screen distance and then dynamically scales a Snellen chart (Sloan or Classic) to the correct physical size.
- Screen calibration: Use a standard credit/debit card to calibrate pixels per millimetre.
- Automatic eye detection: Detects face/eyes using OpenCV YuNet.
- Distance estimation: Computes eye -> screen distance via interpupillary distance (IPD) calibration.
- Dynamic Snellen chart: Supports Sloan, Classic Snellen, or Single letter styles.
- Customisable chart: Polarity, labels, letter spacing, etc.
- Webcam integration: Take a snapshot directly inside the app for calibration.
This project is for educational and research purposes only. It is not a medical device and should not be used as a substitute for professional eye exams.
git clone https://github.com/AnishrKhadka/acuitycheck.git
cd acuitycheck chmod +x ./scripts/download_models.sh
./scripts/download_models.shThis will fetch:
face_detection_yunet_2023mar(face detector, ONNX)
Run the app with:
docker compose up --build Then open your browser at http://localhost:8501.
- Card calibration → Align the on-screen rectangle with your real credit/debit card.
- Snapshot → Take a webcam photo. The app detects your eyes and estimates pixel IPD.
- Calibration → Enter a known camera→eye distance once to compute focal length.
- Chart → View a dynamically sized Snellen chart at the centre of the screen.
Pull requests and suggestions are welcome. Please open an issue first for discussion.
MIT License © 2025 AcuityCheck