Kaspi Bank in Kazakhstan has rolled out facial recognition biometric technology at the ATM to provide its customers with the very highest levels of security. KAL, Kaspi Bank’s ATM software supplier, believes that it is one of the first ATM deployments in the world supporting this technology.
The use of facial recognition enables Kaspi to deliver more advanced transactions at its ATMs because facial recognition provides an extra layer of security for both the cardholder and the bank.
Kaspi customers can use their national ID card at the ATM to access all of their Kaspi accounts, rather than being limited to only those accounts associated with their ATM card. With facial recognition, Kaspi Bank can now guarantee its customers the most rigorous security along with the convenience of open banking. For example, bank customers can take out a loan at the ATM using only their mobile phone and a QR code, with facial recognition acting as an additional method of identification.
The bank carried out its first transaction using facial recognition in April 2018. So far, 750 of its 1200 ATMs offer this form of authentication – with the remaining ATMs set to follow suit soon.
The process used by the bank is to capture the facial images of their customers in the bank branch and then store the images in a secure biometric database. When the customer taps their ID card or inserts their bank card at the ATM, KAL software activates biometric software supplied by OpenCV. The system captures multiple facial images and determines the best image to be used for recognition. In the event that none are deemed to be suitable, the customer will be prompted to move closer, step back, remove their hat or sunglasses or whatever is needed to capture a suitable image.
The captured image is sent to the biometric processing system where it is compared to the customer’s image stored in the bank’s ID database. Once verified, the customer can access their accounts and perform authorized transactions.
There are numerous anti-fraud measures built into the system to add to its security. One such measure is the support of an infra-red lens to capture additional facial details to prevent fraud attempts.