ATM Cardless Transaction using Face Recognition

Most of ATMs (Automatic Trading Machine) are, especially in Indonesia, using traditional two-factor authentications. The factors are what we have (ATM card) and what we know (PIN code). This method has proved valid, workable, and useful over a long span of time. But like any other methods, this have weaknesses. If we lost our card or forget our PIN, we are locked from the system unless we contact the bank and provide other credentials to prove that we are us.

In the same time, companies around the world are embracing the face recognition technology. It’s application is wide. Ranging from terrorists or security targets identification in airports, fraud identifications, to something overkill like unlocking your smartphone using your face. Can we implement this technology to make people experience in using ATM better?

Why Two-Factor Authentications?

One-factor authentication is simply not secure. Imagine if the bank only provides ATM card without PIN, when someone stole your card all the money in the bank is theirs. Therefore banks require us to have the combination of both the card and PIN code.

Face over ATM card

Card system has weakness such as easy to forget and can be stolen. We will never left our face inside the machine like people sometimes do with the cards, right? We always bring our face anywhere, with consciousness or not. Thus, it eliminates one weakness card has: availability.

The other thing came in mind is the card counterfeiting. Card duplication is a common modus operandi in Indonesia. The thief put magnetic reader on the machine by substituting the card slot cover. Then, they put keypad cover on top of the real machine’s keypad. Therefore, they can read the PIN code and the same time reading the inserted card’s internal data. By using the data gathered from original card, they create the duplicate.

We cannot duplicate our face, so the above method doesn’t work. But we have a range of similarity with other people. 0 is totally different and 1 is identical twins. Everything between 0 and 1 is the percentage of face features similarity. So the challenge is to reduce the bias and increasing the accuracy of face recognition system.

How about twins? That’s why we always need at least two-factor authentication.


To increase the security of face-recognition based authentication, banks can combine it with other biometrics such as fingerprint or with conventional methods like PIN code. We need to calculate the risk thoroughly and more detailed to find what is the perfect combination to make our ATM experience better. This article is written without regarding the utopia of cashless society because that may takes a very long time to be fully adapted around the world. But at least, banks have options to embrace biometrics for their security measures. And face recognition, with the current pace of development, may provide the answer.