Honest Numbers, Honest Limits
Camera-based estimation is real science with real limits. NOZA's approach: score every scan's quality, publish confidence with every value, refuse to show numbers we can't stand behind — and keep validating.
Confidence on every value
Each measurement reports how reliable this scan was — driven by pulse-signal strength, lighting, motion and face coverage.
Hard quality gates
Captures with too much motion or too little face coverage are rejected before analysis and ask for a rescan.
Reference-device comparisons
We run side-by-side trials against smartwatches on real volunteers to tune the pipeline. Formal clinical-grade validation studies are in development.
NOZA Research
Scientific validation, bias and fairness testing across skin tones and lighting conditions, and research partnerships — the workstream that will publish our evidence.
Known limitations — stated plainly
- Estimates degrade in poor light, with heavy motion, or when the face is partially covered.
- SpO₂ is a relative trend, not a medical-grade oximeter reading.
- Facial BMI, appearance age and lifestyle-risk are appearance-based indicators — useful for trends, never definitive claims about a person.
- No camera-based system replaces clinical measurement. For medical concerns, see a professional with clinical instruments.
Judge it on your own face.
Run a scan, check the confidence score, compare with your own devices.
Run a Scan