In-browser pose tracking
MediaPipe detects 33 body landmarks in real time on the phone, processed on the user's device with GPU/CPU fallback.
All work
Thai CV Risk AR
Live system · WebAR & Health Screening
A mobile Web AR app that uses the phone camera for pose tracking to estimate height and waist circumference, then calculates the 10-year risk of cardiovascular disease with the waist-based Thai CV Risk Score. It runs in the mobile browser with no app to install, and a detailed mode lets users enter extra clinical values.
Visit the live systemScreens from the live system
All screens use Demo Mode, which shows a sample body outline instead of a live camera feed.
Medical disclaimer before use
Body scan — measures height and waist
Risk gauge and recommendations
High-accuracy assessment mode
The challenge
- Cardiovascular risk screening needs tape measures, forms and staff, which makes it slow to run at scale.
- Most people don't know their waist-to-height ratio (WHtR) or where their risk level stands.
- Measuring the body with a single phone camera gives values that shift with distance, camera angle and lighting.
- It must be clear that this is preliminary screening, not a medical diagnosis.
What we built
Height and waist from the camera
Estimates height from head to ankle and waist from an ellipse model, with smoothing to keep the numbers steady.
AR HUD
A result card follows the person on screen, showing risk %, WHtR and risk level instantly.
High-accuracy assessment mode
Starts from the camera values, then adds systolic blood pressure (SBP), diabetes and smoking, shown on a 4-level gauge.
Personalised advice
Lifestyle recommendations based on the risk factors found.
Demo Mode & disclaimer
A simulated mode for demos without the camera, and a medical disclaimer page before use.
Technology
Frontend
AR & Vision
Deploy