Method & evidence
How SkinProof decides that a change is real — and tests you can run to check it yourself instead of taking our word for it.
1
Measure the wobble
Two photos of the same face a minute apart never score the same. The calibration pair measures how much each of your 15 concerns moves on its own.
2
Draw a 95% band
From the paired differences d: band = 1.96 · √(Σd² / n). A change between two scans that is inside the band is indistinguishable from photo-to-photo noise.
3
Only report what leaves it
Every change is expressed in bands: 0.8× is noise, 1.5× is likely real, 2×+ is clear. Trials need 14+ days and average the two latest check-ins.
Limits we're honest about
Fairness across skin tones
Noise is measured per person, so a band that is wider for one face is never applied to another. We haven't yet measured whether YouCam's scores are equally stable across skin tones — the no-change test above is how anyone can check it for their own skin.Light is the biggest source of noise
The camera warns when lighting differs from your calibration photos by more than 20%. Same spot, same time of day, no flash.One calibration pair is a rough estimate
A band from a single pair can be too tight or too wide. The no-change test catches a band that is too tight; adding pairs fixes it.Not a medical device
Scores describe how skin looks in a photo, not its health. See our privacy & safety notes.