Enterprise biometric identification powered by Multi-Layer Anti-Spoofing (EAR blink, 2D FFT moiré, 3D pose), Apple Face ID multi-angle wizard, FastAPI edge microservices, and RTSP IP streaming ✨
Comprehensive 2026 biometric AI architecture and security suite
Dynamic EAR blink detection, 2D FFT screen moiré replay defense, and 3D solvePnP Euler angle head-pose verification with nose ray tracking.
Interactive 36-segment circular HUD guiding Frontal, Left, Right, Up, and Down angles to build robust L2-normalized Fréchet centroid vectors.
SciPy cKDTree Euclidean nearest-neighbor index delivering sub-millisecond recognition across 10,000+ faces at 60+ FPS via YuNet DNN.
High-throughput asynchronous REST API and real-time WebSockets with interactive OpenAPI Swagger UI (/docs) and ReDoc (/redoc).
Threaded IP camera stream with zero-latency buffer-drop and OpenCV MOG2 motion gating reducing idle CPU power consumption by over 90%.
Encrypted biometric templates (.dat), path traversal sanitization, and tamper-evident SHA-256 chained cryptographic audit trails.
Integrated AudioAnnouncer speech engine providing natural voice greetings ("Access Granted, Welcome Nilay!") with zero cloud dependencies.
Automated attendance tracking with Indian Standard Time (IST) logging, exportable to Apple-styled executive HTML dashboards and CSV.
Instant asynchronous HTTP dispatch for access-control IoT door relays, Slack alerts, and Discord channel security notifications.
Complete pytest test suite with 100% passing automated verification across API, Face ID wizard, RTSP streaming, and encryption.
Instant execution scripts: run_gui.bat, run_api.bat, run_face_id_wizard.bat, and run_rtsp_stream.bat for plug-and-play operation.
Interactive anti-spoofing challenge prompts (random blinks, directional head turns) defeating pre-recorded video injection attacks.
Explore the desktop app, enrollment wizard, REST API, and RTSP stream
Real-time video canvas with color-coded bounding boxes, live anti-spoofing feedback, and voice announcements.
Screenshot: FRC SS.png
Interactive 36-segment circular HUD with 5-pose angular calibration (Yaw ±25°, Pitch ±18°) and Frechet centroid vectors.
run_face_id_wizard.batProduction-ready async REST API with interactive Swagger UI (/docs) and WebSocket video stream broadcasting.
/docsrun_api.bat (Port 8000)Threaded zero-latency buffer drop with OpenCV MOG2 background motion gating, slashing idle CPU by over 90%.
[INFO] RTSP Stream initialized: rtsp://192.168.1.100:554/live
[INFO] cv2.CAP_PROP_BUFFERSIZE set to 1 (Zero-Latency Frame Buffer)
[POWER] MOG2 Motion Gating: NO MOTION detected (Idle power-saving mode active)
[EVENT] Motion threshold exceeded (Area: 1,420px) -> Awakening DNN recognition pipeline!
[WEBHOOK] Dispatching HTTP POST to https://company.internal/security/webhook
[SUCCESS] Webhook delivered: 200 OK (Latency: 14ms)
run_rtsp_stream.batPre-configured batch scripts and commands for instant plug-and-play execution.
$ pytest tests/ -v
============================== test session starts ==============================
collected 74 items
tests/test_api.py .................. [ 24%]
tests/test_composite_enrollment.py ......... [ 36%]
tests/test_rtsp_webhooks.py ........ [ 47%]
tests/test_face_recognition.py ................................ [ 90%]
tests/test_security_audit.py ....... [100%]
========================= 74 passed in 2.08s =========================
--api, --port, --host)Choose your preferred launch or installation method
Instant plug-and-play batch scripts for Windows
Clone the repository and run automated setup
git clone https://github.com/AmiNilay/face-recognition-system-version-control.git
cd face-recognition-system-version-control
./INSTALL.bat
Install full dependencies and run verification
pip install -r requirements.txt
pytest tests/ -v
Guides, API specifications, and technical reference
Get up and running with one-click scripts
→Interactive Swagger & REST endpoints
→Complete 2026 enterprise architecture
→Liveness, RTSP, and recognition tuning
→Automated test suite & verification
→Troubleshooting and community help
→Have questions or enterprise feature requests? We'd love to hear from you