Fraud Prevention in ATM using Face Recognition

Abstract

This report is about advanced and alluring techniques used to identify a person from a video frame and report if the person standing in front of the camera is the authentic and authorized user to perform the operation intended on the ATM. Face detection is widely used to identify person and his/her face now a days. It is used almost in every day interacting objects such as smart phones, laptops, digital cameras, many other electronic devices. This report is about usage of deep learning algorithms which can be applied on the current and conventional systems without much hardware or software change. The procedure of integrating a machine learning is explained later in this report. The video feed is directly fed to the deep learning program which resides on either the host (ATM) or a dedicated computing server which then identifies the person who is standing in front of the ATM and triggers respective alerts.