E-payment Systems Security Solutions Using Facial Authentication Based on Artificial Neural Networks

This article discusses the issue of implementing a holistic solution to ensure the security of users’ work in payment systems, which can increase confidentiality and security using the developing technology of biometric face authentication. Biometric authentication methods are the latest and most popular technologies that solve two problems at once: they eliminate the growing security risk associated with using only roles, and are much more user-friendly than a password that is often forgotten. The overall accuracy of the proposed model is 87.5% for face authentication, where the model randomly requests photos based on one of the facial expressions. In this article we will analyze and do the following steps: Image capturing, Image preprocessing, building neural network to detect face, building neural network to recognize face, building neural network to compare recognized face and Result of Identification. The facial expression includes 7 categories: sad, angry, surprise, fear, happiness, disgust, neutral. As a result, the biometric authentication module can be extended to applications such as shopping malls, offices, etc., and should not be limited to only payment systems.