Enhancing E-Banking Security and Personalization through Convolutional Neural Network-Based Facial Recognition

Abstract

Facial recognition technology has become a cornerstone in enhancing the security and user experience of electronic banking services. Its ability to provide a seamless yet secure authentication method is vital in combating fraud and maintaining the integrity of online financial transactions. This study presents the development of an advanced electronic banking system that strengthens security through the integration of a Convolutional Neural Network (CNN)–based facial recognition model. The system combines facial recognition with One-Time Password (OTP) verification to provide multi-factor authentication. The proposed system includes a web-based user interface and is trained to distinguish between genuine and fraudulent facial recognition attempts. The training dataset consists of 26 genuine and 26 fraudulent videos, from which 150 frames were extracted per video, yielding a total of 7,800 frames. The dataset was divided into 80% for training and 20% for testing. The CNN model achieved an accuracy of 99.92% in differentiating between real and spoofed facial images, demonstrating strong capability in detecting attempts to deceive the system. This high accuracy highlights the model’s robustness and reliability in securing electronic banking services, significantly enhancing the safety and integrity of online financial transactions. The integration of OTP verification provides an additional security layer, ensuring that even if facial recognition were compromised, unauthorized access would still be prevented. Overall, the results emphasize the potential of deep learning techniques and multi-factor authentication in strengthening cybersecurity measures within the banking sector.