IoT-Based Biometric Access Control System
2026 IEEE Zooming Innovation in Consumer Technologies International Conference, ZINC 2026, Novi-Sad, Serbia, 3 - 04 June 2026, pp.135-140, (Full Text)
- Publication Type: Conference Paper / Full Text
- Doi Number: 10.1109/zinc69910.2026.11655703
- City: Novi-Sad
- Country: Serbia
- Page Numbers: pp.135-140
- Keywords: Biometric Access Control, Edge Computing, Face Recognition, HMAC-SHA256, Internet of Things, Multimodal Biometrics
- Dokuz Eylül University Affiliated: Yes
Abstract
The traditional biometric access systems have significant weaknesses in providing the present day's security standards. Two of the weaknesses are data interception and presentation attacks. Our study aims to present a low-cost IoT-based biometric security tool, as the first phase of comprehensive secure access systems. The proposed edge device architecture utilizes an AIThinker ESP32-CAM edge device, an OV2640 Camera, and an HC-SR04 Ultrasonic distance sensor as the hardware. To protect the data integrity and mitigate man-in-the-middle attacks, an HMAC-SHA256 Cryptographic Layer is integrated. The validated image data is then processed by a Flask-based local Python server, which computes RetinaFace for accurate face detection, and Facenet512 for face feature extraction. The identification process is completed via a cosine similarity calculation inside the ChromaDB vector database. The process is asynchronously logged to the Cloudinary platform. Our work presents the first phase of a multimodal biometric security framework that can be integrated with EEG signals to ensure liveness detection and prevent spoofing attacks.