A biometric-based authentication system for e-security

• Main Author: Tharuka, Garagoda Arachchige Wageesha Nirmal
• Other Authors: Mohammed Yakoob Siyal
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Engineering; biometric
• Online Access: https://hdl.handle.net/10356/176180
• Institution: Nanyang Technological University

The report explores into the realm of cybersecurity and authentication methods, particularly focusing on the implementation of biometric authentication systems, specifically static biometrics. It provides an in-depth analysis of technical security measures and human and organisational factors crucial for fortifying digital infrastructures against potential threats. The exploration of authentication mechanisms encompasses knowledge-based, possession-based, and biometric authentication, with a special emphasis on the principles and comparative evaluation of various biometric systems. Face recognition principles are thoroughly examined, elucidating face detection techniques utilising Haar-cascade classifiers, feature extraction methods like Principal Component Analysis, and feature classification utilising Support Vector Machines. SVM are highlighted as key components within face recognition systems. These algorithms play a crucial role in accurately categorising facial features extracted from images. SVMs excel in their ability to discern between different individuals by identifying the optimal hyperplane that maximises the margin between classes, thus enhancing the system's generalisation capability. Additionally, SVMs are well-suited for handling high-dimensional data, making them ideal for face recognition tasks where numerous facial features are considered simultaneously. Through the integration of SVMs, face recognition systems can achieve robust performance and adaptability across diverse datasets and environmental conditions. A detailed implementation strategy is outlined, delineating a multi-stage process for developing a real-time face recognition system. This process includes comprehensive steps such as data collection, preprocessing, augmentation, and the application of advanced algorithms. The report further scrutinises the results of accuracy and reliability testing, considering variables such as lighting conditions and facial expressions to ensure robust performance. In conclusion, the report underscores the continuous endeavour to enhance the accuracy and reliability of face recognition systems. It also proposes future directions for research and implementation, including the integration of databases and the deployment of systems online. This comprehensive examination of cybersecurity and authentication methods serves as a foundational guide for organisations aiming to bolster their defences against evolving digital threats while leveraging cutting-edge biometric technologies for enhanced security measures.