AI Facial Recognition Smart Locker Security System with Multi-Level Authentication

Abstract

AI Facial Recognition Smart Locker Security System with Multi-Level Authentication is an intelligent AI-based security platform designed to provide secure, reliable, and keyless access control for lockers and restricted areas through multi-factor authentication techniques. In today\'s digital era, traditional locker systems that rely on physical keys, passwords, or RFID cards are vulnerable to theft, duplication, unauthorized access, and security breaches. This project addresses these limitations by offering a real-time, intelligent, and multi-level authentication mechanism using Artificial Intelligence and Machine Learning techniques. The system analyzes user authentication inputs such as facial features, behavioural patterns, voice commands, and One-Time Password (OTP) verification to accurately identify authorized users and prevent unauthorized access. It integrates modern technologies including Artificial Intelligence, Computer Vision, Machine Learning algorithms, biometric authentication methods, email-based OTP services, and database management systems to ensure security, efficiency, scalability, and real-time monitoring. By combining facial recognition intelligence with multi-level authentication and intruder detection mechanisms, the AI Facial Recognition Smart Locker Security System aims to provide a secure, intelligent, and scalable access control solution suitable for banks, offices, smart vaults, and other restricted environments.

Introduction

Security systems are essential for protecting valuable assets and personal information, but traditional locker systems based on physical keys, passwords, and RFID cards face major vulnerabilities such as theft, duplication, hacking, and unauthorized access. To overcome these limitations, the proposed AI Facial Recognition Smart Locker Security System with Multi-Level Authentication uses Artificial Intelligence, Machine Learning, and Computer Vision to provide secure, intelligent, and real-time access control.

The system identifies authorized users through facial recognition and enhances security using additional authentication layers, including behaviour authentication, voice authentication, OTP verification, and PIN backup. This multi-level approach improves authentication accuracy, prevents unauthorized access, and enables real-time monitoring and alerts.

Existing System and Limitations

Traditional locker security systems mainly depend on single authentication methods such as keys, passwords, or RFID cards. These methods have several weaknesses:

• Physical keys can be lost, stolen, or duplicated.
• Passwords can be shared, guessed, or hacked.
• RFID cards can be cloned or misused.
• Lack of biometric verification makes accurate user identification difficult.
• Absence of real-time monitoring and alerts reduces security effectiveness.
• Limited data management capabilities prevent scalability and intelligent security management.

Proposed System

The proposed AI-based locker system provides a reliable and scalable security solution by integrating:

• Facial Recognition: Identifies users by comparing facial features with stored biometric data.
• Behaviour Authentication: Analyses user activity patterns for additional verification.
• Voice Authentication: Uses voice characteristics for identity confirmation.
• OTP Verification: Provides an additional security layer through one-time passwords.
• Email Alerts: Sends notifications during authentication events or suspicious activities.
• Database Management: Securely stores user details, biometric information, and security records.

Algorithms Used

The system applies Artificial Intelligence and Machine Learning algorithms to improve authentication performance:

• Decision Process: Evaluates biometric inputs and determines whether access should be granted or denied.
• Error Function: Measures authentication accuracy by comparing captured data with stored information.
• Optimization Process: Improves model performance by reducing errors and updating parameters.
• Supervised Learning: Uses labeled biometric data for accurate user identification.
• Unsupervised Learning: Detects hidden patterns and unusual activities.
• Semi-Supervised Learning: Combines labeled and unlabeled data to improve learning efficiency.

Techniques Implemented

The system uses advanced techniques to provide secure and efficient operation:

1. Facial Recognition: Enables accurate biometric identification.
2. Multi-Level Authentication: Combines multiple verification methods for stronger security.
3. Email Integration: Supports OTP delivery and security notifications.
4. Database Management: Maintains organized and secure storage of user and authentication data.

Methodology

The system follows a structured authentication process:

1. Input Collection: Users provide facial images, voice samples, or behavioural information through the interface.
2. Data Processing: AI and Computer Vision techniques extract relevant features from biometric inputs.
3. Verification: The processed information is compared with stored biometric records.
4. Decision Making: Access is granted only when authentication requirements are successfully satisfied.

Conclusion

SmartLocker AI provides a secure and intelligent authentication system using facial recognition, voice verification, behavioural biometrics, and email OTP. The system enhances security, reduces unauthorized access, and offers efficient user authentication. Overall, it delivers a reliable and modern smart locker security solution.

References

[1] Viola, P., & Jones, M. (2001). “Rapid Object Detection using a Boosted Cascade of Simple Features.” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 1, 511–518. [2] Turk, M., & Pentland, A. (1991). “Eigenfaces for Recognition.” Journal of Cognitive Neuroscience, 3(1), 71–86. [3] Schroff, F., Kalenichenko, D., & Philbin, J. (2015). “FaceNet: A Unified Embedding for Face Recognition and Clustering.” Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 815–823. [4] Zhang, K., Zhang, Z., Li, Z., & Qiao, Y. (2016). “Joint Face Detection and Alignment Using Multi-task Cascaded Convolutional Networks.” IEEE Signal Processing Letters, 23(10), 1499–1503. [5] Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press. [6] Bishop, C. M. (2006). Pattern Recognition and Machine Learning. Springer. [7] Bradski, G. (2000). “The OpenCV Library.” Dr. Dobb\'s Journal of Software Tools. [8] Zeng, W., & Zhao, Y. (2022). “Artificial Intelligence-Based Smart Security Systems: A Survey.” Journal of Information Security and Applications, 65, 103091. [9] Jain, A. K., Ross, A., & Nandakumar, K. (2011). Introduction to Biometrics. Springer. [10] Daugman, J. (2004). “How Iris Recognition Works.” IEEE Transactions on Circuits and Systems for Video Technology, 14(1), 21–30. [11] Rabiner, L. R., & Juang, B. H. (1993). Fundamentals of Speech Recognition. Prentice Hall. [12] Howard, A. G., et al. (2017). “MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications.” arXiv preprint arXiv:1704.04861. [13] Szeliski, R. (2022). Computer Vision: Algorithms and Applications (2nd ed.). Springer. [14] Stallings, W. (2017). Cryptography and Network Security: Principles and Practice (7th ed.). Pearson. [15] Sandhu, R., & Samarati, P. (1994). “Access Control: Principle and Practice.” IEEE Communications Magazine, 32(9), 40–48.

Copyright

Copyright © 2026 M. Nagakeerthi, Etamsetti Padma. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.