Violence Detection System 🚨

A web-based AI system to detect violent events (fights, fire, crashes) in videos using CLIP-based models with real-time alerts.

---

🌟 Features

• 📂 Upload and analyze videos (up to 70MB)
• ✨ Detects violence using OpenAI CLIP model
• 📹 Real-time feed with violence overlay
• 🚨 Visual & sound alarms when violence is detected
• 📢 Browser notifications
• ♻️ Motion detection optimization
• 🌐 Responsive modern web UI

---

📁 Project Structure

CV/
  cv/
    app.py                # Flask app
    assets/
      demo.gif            # Demo animation
    static/
      css/
        style.css         # Styles
      js/
        script.js         # Frontend logic
    templates/
      index.html          # Web page

---

🚀 Requirements

• Python 3.8+
• pip

📊 Python Dependencies

• Flask
• opencv-python
• numpy
• torch
• transformers
• pillow
• werkzeug

---

🔧 Usage

1. Run the server

python app.py

2. Visit in your browser

http://127.0.0.1:5000/

3. Upload a video

• Max size: 70MB
• Wait for analysis to complete

4. Watch for alerts

• Overlay text
• Sound + visual warnings
• Notification popup

---

Technical Highlights

CLIP Model by OpenAI is used for zero-shot classification.

Motion Detection reduces unnecessary frame classification.

Frontend Alerts via JS + HTML5 Audio + Notification API.

Stateless Design: Suitable for further scaling and microservice use.

✅ Future enhancements

🔐 Role-based access

☁️ Cloud storage

📈 Admin dashboards

🚔 Law enforcement integration

🎥 Multi-camera support

✨ Made with passion by Raghuraam ✨