The science of AverScan
How a camera reads your pulse from the color of your skin — the genuinely cool part.
This is the page where AverCare goes from "neat app" to "wait, how?" Reading vitals from a plain video of someone's face sounds like science fiction, but it's real, well-studied physiology plus some clever signal processing.
The core idea: your face blushes with every heartbeat
Every time your heart beats, it pushes a pulse of blood through the vessels just under your skin. That extra blood absorbs a tiny bit more light, so your skin gets imperceptibly redder and then fades back — dozens of times a minute. You can't see it. A camera, looking at many pixels over many frames, can.
It has a name
Measuring blood-volume changes from light is called photoplethysmography (PPG). The pulse oximeter on your finger at the doctor's office does it with an LED pressed to your skin. Doing it from a distance with a camera is remote PPG — rPPG.
From pixels to a pulse
Capture clean frames
The app records a short video of your face. Quality gates check lighting, framing, and stillness — garbage in would mean garbage out.
Find the skin
Face-landmark detection isolates the regions that carry the best signal (like the forehead and cheeks) and ignores eyes, hair, and background.
Watch the color shift
Across the frames, the app tracks minuscule changes in skin color over time. Plot that and you get a wavy line — a waveform.
Find the rhythm
The repeating peaks in that waveform are your heartbeats. Their spacing gives heart rate; subtle variation in the spacing gives heart-rate variability; the shape carries even more.
Why stillness and light matter so much
The signal AverScan is chasing is tiny — a fraction of a percent change in color. Motion and flickering light create changes far bigger than the heartbeat itself, which can swamp it. That's why the app insists on good conditions before it trusts a reading. It's not being fussy; it's protecting the result.
Heart-rate variability, briefly
A healthy heart doesn't tick like a metronome — the gaps between beats vary a little, and more variation is generally a good sign of a relaxed, adaptable nervous system. That variation is also how AverCare can estimate a stress indicator without a single extra sensor.
Two engines, one result
AverCare supports two scan modes (see AverScan): an on-device analysis that works anywhere, and an opt-in research-grade engine built on transdermal optical imaging for a richer panel of readings. Both feed the same friendly results screen, so the experience is identical — only the math behind the curtain differs.
The honest caveats
Camera-based vitals are estimates, and they're best at spotting trends over time rather than nailing a single clinical number. AverScan is a wellness tool, not a medical device. Skin tone, lighting, makeup, and movement all affect accuracy — which is exactly why AverCare is upfront about conditions and never dresses an estimate up as a diagnosis.