Fawry Surveillance System

A FastAPI based full-stack interface for real-time multi-object tracking and face recognition in retail surveillance video streams. Comes with:

• 🎥 Integration Page: run tracking + face recognition in one go
• 🧑‍💻 Face Recognition: upload a staff image, identify the person
• 🚶‍♂️ Object Tracking: upload a video, get an annotated tracking output
• 💾 Database: PostgreSQL
• ⚙️ Modular Services & Routers for easy extension

---

🚀 Quick Start

1. Clone & enter the repo

   git clone https://github.com/your-org/surveillance-system.git
   cd surveillance-system

2. Create a virtual environment

   python -m venv .venv
   source .venv/bin/activate      # macOS/Linux
   .\.venv\Scripts\activate       # Windows PowerShell

3. Install dependencies

   pip install -r requirements.txt

4. Configure environment Copy .env.example → .env and fill in:

   DATABASE_URL=postgresql://user:pass@localhost:5432/dbname
   SECRET_KEY=your_jwt_secret
   HASH_ALGORITHM=HS256
   ACCESS_TOKEN_EXPIRE_MINUTES=60
   TRACKING_MODEL_PATH=/path/to/your/trackingModel.pt

5. Run the application

   uvicorn app.main:app --reload

   Open your browser at http://127.0.0.1:8000

---

🗂️ Project Structure

surveillance-system/
├── app/
│   ├── main.py              # FastAPI app + routes & static mounts
│   ├── core/                # Config, database, security
│   ├── models/              # SQLAlchemy models (User, Branch, Region…)
│   ├── schemas/             # Pydantic schemas for requests/responses
│   ├── services/            # Business logic: tracking, face, integration
│   ├── routers/             # API routers: auth, users, face, tracking, integration
│   ├── templates/           # Jinja2 HTML pages (index, face, tracking, results)
│   └── static/              # Front-end assets
│       ├── css/style.css
│       ├── js/
│       │   ├── main.js      # integration page logic
│       │   ├── face.js      # face-recognition page logic
│       │   ├── tracking.js  # object-tracking page logic
│       │   └── results.js   # shared results-page logic
│       ├── images/          # logos, icons
│       └── outputs/         # auto-served annotated videos
├── requirements.txt
├── .env.example
└── README.md

---

📄 Features & Endpoints

1. Integration (Tracking + Face)

• UI: GET /
• Process: POST /models/integration/video
• Results: redirects to /results, shows annotated video & combined detections

---

2. Face Recognition

• UI: GET /face
• API: POST /models/face/recognize
  • Input: form-file file (image)
  • Output: { "person": "<matched_name>" }

---

3. Object Tracking

• UI: GET /tracking
• API: POST /models/tracking/video
  • Input: form-file file (video)
  • Output:

    {
      "annotated_video": "/static/outputs/xyz_annotated.mp4",
      "predictions": "tracker_results/data/xyz.txt"
    }

---

🔧 Front-End Pages

1. Home / Integration

   • Drag-and-drop or click to select a video
   • Yellow highlight on selection
   • Preview results & download annotated video

2. Face Recognition

   • Upload an image
   • Displays name of matched staff member

3. Object Tracking

   • Upload a raw video
   • Embeds the annotated tracking video

All pages share a professional, responsive design with the Fawry brand colors (yellow & blue) and logo.