A production-oriented face verification pipeline built with deep learning, designed for real-world security applications.

Face Recognition V2 is a real-time face verification system that implements a complete machine learning pipeline — from data acquisition to live identity verification.

Unlike basic demos, this project focuses on system design, automation, and real-world usability, making it a strong foundation for scalable security solutions.

It serves as the core engine for the upcoming USB Locker V3, where facial authentication will control system-level access.

• ⚡ Real-time face detection & verification
• 🧠 Deep learning pipeline (MTCNN + FaceNet)
• 🔄 Fully automated workflow (no manual setup)
• 📸 Live dataset generation from camera
• ⚙️ Precomputed embeddings for fast inference
• 📱 Mobile camera support via DroidCam
• 🧩 Modular, production-style architecture

1. Data Collection (via DroidCam)
2. Face Detection (MTCNN)
3. Embedding Extraction (FaceNet)
4. Real-Time Recognition

This system requires zero manual setup:

• If no dataset → automatically captures images
• If no embeddings → builds them automatically
• If everything exists → runs instantly

Face_Recognition_V2/
│
├── data/
│   ├── raw/
│   └── embeddings/
│
├── src/
│   ├── pipeline/
│   │   ├── detector.py
│   │   ├── embedder.py
│   │   ├── recognizer.py
│   │   ├── capture_dataset.py
│   │   └── droidCam.py
│   │
│   └── main.py

pip install -r requirements.txt
python src/main.py

• Detection: MTCNN
• Embedding: FaceNet (keras-facenet)
• Verification: Euclidean distance + thresholding
• Pipeline: Fully automated (capture → build → run)

• Real-time execution on CPU
• No GPU required
• Efficient and lightweight pipeline
• Stable under different lighting conditions

• Built and tested on a low-end personal machine
• Designed to work under limited resources
• Focused on practical deployment, not just theory

• Secure workstation access
• Face-based authentication systems
• Security prototypes
• Foundation for hardware-integrated solutions

This project is the foundation for USB Locker V3, where face recognition will be integrated with system-level security controls.

ATEF Aliat Machine Learning & Computer Vision Enthusiast

• Focus: Real-time systems, AI pipelines, and security applications
• Built projects under hardware constraints to simulate real-world conditions
• Strong interest in applied AI and system design

This project demonstrates the ability to:

• Design and implement end-to-end ML systems
• Work with real-time data pipelines
• Build practical solutions under constraints

Not just a model — but a complete, working system.