Face recognition + Voice recognition + Document verification

How it works

Customers are requested as either part of a new account registration process or an account update, to add their face, voice and document on a self service basis or via a video chat.

After the customer fills out your organizations standards questions such as name, address date of birth etc, a request appears to add their biometric data. If this is during an enrolment on an app, the camera on the device will turn on and a live stream will start during which the customer chooses to take a picture of themselves. Once a picture has been chosen that the customer is happy with, it then gets sent to the server to confirm it is a unique image.

The customer is then requested to add a voice print. This is a series of numbers randomly generated and takes up to six times of repeated numbers to obtain an accurate print. The voice print is then associated only with the customers face image.

The final part of the process is capturing the data and details from the ID document. This data will then be compared to the data that the customer added and if there are any inconsistencies then the data from the ID document will be added over the data that was inputted manually by the customer. This prevents fraud from taking place and the ability of one document to be used by more than one person.

Strong authentication

By combining a two part authentication with additional document verification it enables a very strong authentication to take place.

No more passwords

Using a two part authentication it will provide the first strong authentication process that will be able to replace passwords. It will now not be necessary to remember your passwords as you will be able to log in to your account simply by using your face and your voice.

Self service accounts

By using mobile devices and PC’s the customer takes on all of the work. The process is conducted through apps on the device or directly on the web page of their account. No human interface is required either on a PC or via a telephone call.

Almost zero false positives

There will be virtually no false positives using a two factor biometric authentication. If false positives take place during the face recognition process and a number of faces are flagged up as looking similar, the voice recognition then compares the voice pattern to those false positives. It will return a result that is likely to show only one match.

Thus false positives will become almost non-existent.

Use cases

One of our customers requested that we set up a KYC process that included face, voice and document verification for mobile devices and PC’s. Once registered the customer could then access their account using either their face, face and voice or password depending on the level of security required for the transaction.

At each stage of the process there was always a fall back plan so if the customers camera was not working then they could choose voice recognition for example. If their voice recognition did not work, then they could choose video chat. If nothing worked then they could use the old fashioned method and call up the call centre and confirm their identity that way using passive voice recognition. But the customer had to have the ability to always access their account.