Facial Authenticity and Spoofing Detection

The growing use of biometric authentication, specifically facial recognition software, in today’s digital world has emphasized the need for strong security measures to block unauthorized access attempts. There has been a major reliance on this type of authentication method recently, our study addresses these major challenges related to detecting authentic facial features and averting identity fraud through these different spoofing techniques. By conducting thorough research as well as examining these various types of spoof attacks that are putting face recognition systems at risk, we can argue that there is a need for advanced detection mechanisms. By going through and incorporating previous research and methodologies into our approach, we propose a new approach by utilizing Local Binary Patterns (LBP) and an orchestra of classifiers comprised of Random Forest, K-Nearest Neighbors, and XGBoost algorithms. We conducted substantial tests with a wide variety of datasets which showed promising recall performance along with accuracy and precision when comparing between real and fraudulent faces. Increasing the overall effectiveness at reducing flaws inherent in face recognition platforms while enhancing the user trust and reliability on biometric identification methods.