Face verification

What it does

Face verification takes the selfie from the session and compares it with the portrait on the document. It runs in every verification, right after the document step and alongside liveness.

Checks

One-to-one face match between the selfie and the document portrait.
Foreign-face detection.
Age check against the date of birth read from the document, with a configurable age window of 16 to 100.
Deepfake and face-swap detection inside the selfie and liveness steps.
Face attributes such as age, gender, emotion, glasses, mask and head pose.

Add-ons you can switch on

Repeat-user search and blocklists find returning faces among earlier verified users and check your own person lists and blocklists.
Re-verification of a returning user compares a returning user with a stored reference session.

How it works in the flow

The user takes a selfie in the widget.
The selfie is compared with the document portrait.
The face match result is added to the session.

What your system receives

A selfie entry in the session results, with the status of the step.

Face API

You can also call detect, compare and search on their own, without a full session, through the face API. Read the REST API overview.

Standards

Our face matching algorithm reports TOP-24, NIST FRVT 1:1 Verification. This is vendor-reported. See standards and statements.

Questions, answered plainly

What does face verification compare?

The selfie taken in the session against the portrait on the document. It is a one-to-one comparison.

Can I set an age limit?

Yes. The age check compares the date of birth on the document with your age window, which is configurable from 16 to 100.

Does it detect deepfakes?

Deepfake and face-swap detection runs inside the selfie and liveness steps and spots generated and swapped faces.

Can I find returning users?

Yes. Repeat-user search finds a returning face among earlier verified users, and your own person lists and blocklists can be checked in the same flow.