Smart Door — Face Recognition Access Control
AI-powered smart door system using ESP32-CAM, Raspberry Pi, FaceNet, and MTCNN with liveness detection.

Overview
An AI-powered smart door access control system that uses face recognition to allow or deny entry. The system employs an ESP32-CAM module for image capture, a Raspberry Pi for edge processing, FaceNet for face embedding and matching, MTCNN for face detection, and a texture-based liveness detection algorithm to prevent spoofing attacks with photos or screens. Only known/registered faces are granted access.
The system includes cloud integration for remote monitoring, real-time notifications, and an admin interface for managing authorized users. Currently an active hobby project under continued development.
Key Features
Technical Highlights
Combined deep learning (FaceNet + MTCNN) with embedded hardware
Implemented anti-spoofing liveness checks at the edge
Designed low-latency pipeline for real-time door access decisions