Our in-house processing engine uses remote photoplethysmography (rPPG) to turn any camera into a source of vital signs: heart rate, HRV, breathing and blood pressure, no cuff, no wearable, in a 30-second face scan.
Every heartbeat changes the blood volume in your face, and with it the colour of your skin: invisible to the eye, measurable by a camera. That is remote photoplethysmography (rPPG), and we built our processing engine around it, in-house.
Inside our health sensing layer, every scan passes six stages (frame capture, frame quality check, signal extraction, signal quality check, biomarker extraction and biomarker quality check), built on three proprietary technologies: TripleGate, VitalMesh and PulseShape.
Data points used to test and validate our models.
Research papers by our R&D team, in venues such as CVPR Workshops, BMVC, IEEE T-BIOM and Frontiers in Digital Health.
Countries where our engine is live.
Three quality gates, at frame, signal and biomarker level.
No value is released unless all three gates pass.
Our own skin modelling finds the skin regions that carry a usable pulse, on the forehead and cheeks, and tracks them through the scan while compensating for movement. Our colour-projection method then turns those pixels into a pulse signal.
Breathing is measured in parallel from chest motion, with in-house segmentation and keypoint networks. Several independent methods have to agree before a breathing signal counts.
PulseShape measures blood pressure without a cuff: it reads the shape of the pulse wave with a feature extraction model and our proprietary temporal model, and combines it with context such as age and sex.
Following a standardised protocol, every biomarker is tested against medical reference devices.
of measurements within 5 bpm of the reference. RMSE 1.67 bpm.
of measurements within 3 breaths per minute of the reference. RMSE 1.83 brpm.
mean error ± standard deviation compared with the reference.
of measurements within 40 ms of an ECG chest belt. Mean absolute error 13 ± 15 ms.
Whether you build a care app, a kiosk or a care robot, the scan fits into moments your users already have. Let’s find yours.
Why wait? Let’s start a conversation.
