Face recognition works via the inbuilt camera on the device. As soon as a request is made to capture the customer’s face for verification purposes, the camera on their device turns on and the customer is requested to agree to an image of their face being taken.
The facial image is then compared on a 1:N (one to many) basis to see if the face is already within the database. If it is then potentially an account fraud is taking place.
Once the facial image has been added to a customers account the customer can use their facial biometric to log into their account.
Depending on the level of risk associated with the login, the image is either compared on the device for a low risk authentication or on the server for a high risk verification.
A liveness test will also be conducted that requires the customer to make a movement, such as blinking, which tells the software that the person is real.
This liveness detection is a low level one compared to the combination of face and voice liveness detection and should not be used for high security logins.
When combined with other checks such as voice recognition and document verification facial recognition becomes a very useful tool to help with KYC and anti fraud issues.
The whole process can be conducted automatically so that the customer self serves on any device.
It is possible to add a face on an android device and immediately use that face on a PC to log in to an account. The face data is cross channel.