In the current digital era, where cybersecurity and data privacy are paramount, the limitations of traditional authentication mechanisms such as passwords, PINs, and OTPs have become evident. These methods are prone to vulnerabilities including credential theft, phishing attacks, and poor user compliance. To address these challenges, biometric authentication has emerged as a secure and user-centric solution, with facial recognition gaining prominence due to its nonintrusiveness, accuracy, and ease of integration. This paper presents the design and implementation of a Deep Learning–based Face Recognition Login System, which eliminates the reliance on conventional credentials by leveraging real-time biometric verification. The system integrates OpenCV for face detection, a pre-trained OpenFace model for feature embedding, Flask for backend processing, and MongoDB for secure storage of user embeddings and profiles. A responsive web interface (HTML, Tailwind CSS, JavaScript) facilitates user registration, login, and profile management, while robust security mechanisms ensure encrypted storage and session integrity. The proposed system demonstrates high reliability in real-time conditions with varying lighting and positional challenges, offering superior usability compared to existing password-based systems. Beyond authentication, the framework establishes a foundation for broader applications such as attendance management, role-based access control, and integration with enterprise security systems. This research thus contributes to bridging the gap between academic deep learning models and practical, deployable authentication solutions.
Akshay Deepak (Thu,) studied this question.