🎓 AI Face Recognition Movement & Attendance System

An intelligent college entry-exit monitoring system powered by Face Recognition, built using Flask, DeepFace, and SQLite.

This system enables automated attendance tracking, movement logging across multiple campus locations, real-time analytics, and Excel export functionality.

---

🚀 Features

• 🧠 Support for Multiple Face Embeddings (Higher accuracy)
• 📷 Live camera-based registration & Recognition
• 📤 Upload Photo(s) support for registration
• 🗂 High-performance face embedding caching (Pickle-based)
• 🧪 Multi-face recognition support in single frame
• 🔁 Smart entry/exit toggle logic per location
• 🏫 Multi-location support (Gym, Library, Hostel, etc.)
• 📊 Live occupancy dashboard
• 📈 Daily analytics view
• 📥 Excel log export (.xlsx)
• 🔐 Admin & Guest access control
• 💾 Lightweight SQLite database

---

🏗 System Architecture

Camera / Upload
    ↓
Face Encoding (DeepFace)
    ↓
Multi-Embedding Comparison (Max Similarity)
    ↓
Entry / Exit Logging
    ↓
Live Dashboard & Analytics

Core Components

Component	Responsibility
Database Layer	Handles student & movement storage
Face Encoder	Generates embeddings using DeepFace
Recognition Service	Matches faces via max similarity comparison
Movement Manager	Controls entry/exit toggling
Caching System	Fast startup & load using face_cache.pkl
Flask App	Web interface & API layer

---

🛠 Tech Stack

• Flask – Backend framework
• DeepFace – Face recognition engine (Facenet model)
• TensorFlow – Deep learning backend
• OpenCV – Camera integration
• NumPy – Numerical computations
• SQLite3 – Embedded database
• Pandas + OpenPyXL – Excel export
• Pickle – Embedding serialization/caching
• Pillow – Image processing

---

📂 Project Structure

AI-Face-Recognition-System/
│
├── app.py                  # Main Flask application
├── faces.db                # SQLite database (auto-generated)
├── face_cache.pkl          # Performance cache (auto-generated)
├── templates/              # HTML templates
├── static/                 # CSS / JS files
├── requirements.txt
├── README.md
└── .gitignore

---

⚙️ Installation

1️⃣ Clone Repository

git clone https://github.com/Parth844/AI-Face-Recognition-System.git
cd AI-Face-Recognition-System

2️⃣ Create Virtual Environment (Python 3.10 Recommended)

python3.10 -m venv venv
source venv/bin/activate

⚠️ Recommended Python Version: 3.10 or 3.11
TensorFlow / DeepFace may not support latest Python versions.

3️⃣ Install Dependencies

pip install -r requirements.txt

---

▶️ Run Application

python app.py

Open in browser:

http://127.0.0.1:5000

---

👨‍💼 Admin Login (Demo)

Username: admin
Password: password123

⚠️ Change credentials before production use.

---

📊 Capabilities

• Higher Accuracy: Stores multiple embeddings per student.
• Improved Performance: Uses local caching for instant comparisons.
• Smart Logic: Prevents duplicate rapid scans.
• Multi-Location: Tracks movements across different campus gates.
• Excel Reports: Export logs for attendance verification.

---

🔐 Security Notes

• Session-based authentication
• SQLite local storage
• Change SECRET_KEY before deployment
• Avoid committing database in production

---

🔮 Future Improvements

• Face anti-spoofing (liveness detection)
• Role-Based Access Control (RBAC)
• REST API expansion
• WebSocket real-time dashboard
• Docker deployment

---

👨‍💻 Author

Parth Tyagi
AI & Machine Learning Enthusiast
Building intelligent campus automation systems.

---

📜 License

MIT License