Auto Disclaimer 🚭 🍺

📋 Table of Contents

• Overview
• Problem Statement
• Solution
• Demo
• UI Screenshots
• Tech Stack
• Current Challenges
• Future Improvements
• Installation and Usage
  • Docker Deployment
  • Manual Installation
• Usage Instructions
• Performance Tips
• Contact
• License
• Acknowledgements

📝 Overview

Auto Disclaimer is an automated tool that adds smoking disclaimers to movies by detecting smoking scenes. The tool uses computer vision and machine learning to identify frames containing smoking activities and overlays appropriate disclaimers, eliminating the need for manual frame-by-frame editing.

❓ Problem Statement

Movie editors currently spend significant time manually adding disclaimers to movies on a frame-by-frame basis whenever smoking scenes appear. This process is:

• ⏱️ Time-consuming
• ❌ Prone to human error
• 💰 Costly for production companies

💡 Solution

Our solution automates the disclaimer addition process by:

1. Processing movie files frame by frame
2. Using the Moondream2 vision language model to detect smoking activities
3. Automatically overlaying appropriate disclaimers on relevant frames
4. Generating a new video with disclaimers integrated

🎬 Demo

Check out a sample processed video: Example Output

🖼️ UI Screenshots

Main Interface	Processing Results

🛠️ Tech Stack

Technology	Purpose
Python	Core programming language
Moondream2	Vision language model for smoking detection
OpenCV	Movie processing and frame manipulation
Gradio	User-friendly web interface
Docker	Containerization for easy deployment

⚠️ Current Challenges

Challenge	Description
Processing Speed	Takes ~4 minutes to process a 4-minute video at 5 FPS
Model Limitations	Moondream2 (2B parameters) not optimized for smoking detection
Resource Intensive	High computational requirements for full-length movies

🚀 Future Improvements

• 🧠 Develop a custom CNN model specifically trained on smoking detection data
• 🍺 Expand detection capabilities to include drinking scenes
• ⚡ Optimize processing pipeline for better performance

📥 Installation and Usage

Option 1: Docker Deployment (Recommended) 🐳

1. Clone the repository:

   git clone https://github.com/kevinnadar22/AutoDisclaimer.git
   cd AutoDisclaimer

2. Build and run with Docker Compose:

   docker-compose up -d

3. Access the application at http://localhost:7860 in your browser

Option 2: Manual Installation 💻

1. Clone the repository:

   git clone https://github.com/kevinnadar22/AutoDisclaimer.git
   cd AutoDisclaimer

2. Install system dependencies:

   sudo apt-get update && apt-get install -y libvips libgl1 libglib2.0-0

3. Install Python dependencies:

   pip install -r requirements.txt

4. Run the application:

   python gradio_app.py

5. Access the application at http://localhost:7860 in your browser

📋 Usage Instructions

1. 📤 Upload a video file through the web interface
2. Set the frames per second (FPS) for processing (lower values = faster processing but may miss some scenes)
3. Select the detection method (point, detect, or query) - Point is the most accurate
4. Optionally upload a custom disclaimer image
5. ▶️ Click "Process Video" to start the detection and disclaimer addition
6. View the processing statistics and download the processed video when complete

⚡ Performance Tips

Tip	Description
Lower FPS	Use 1-5 FPS for faster processing
GPU Acceleration	Significantly improves processing speed
Movie Splitting	Split long movies into smaller segments
Detection Method	"Point" method is generally faster than others

📞 Contact

• Telegram: t.me/ask_Admin001
• Email: jesikamaraj@gmail.com

📄 License

MIT License

🙏 Acknowledgements

• Moondream2 by Vikhyat K for the efficient vision language model