Cognitive Video AI

Overview

Cognitive Video AI redefines video editing, making it as simple as typing a sentence. Powered by advanced AI, this tool automates video editing tasks based on user prompts, saving time and effort while enabling unmatched creativity. Whether you're a beginner or a professional, Cognitive Video AI bridges the gap between vision and execution.

Features

1. AI-Powered Editing:

   • Automate edits using natural language prompts.
   • Perform tasks like trimming, adjusting brightness, enhancing contrast, sharpening, slowing or speeding up videos, and removing objects.

2. User-Centric Design:

   • Designed for non-experts and professionals alike.
   • Simplifies complex video editing workflows.

3. Seamless Integration:

   • Leverages advanced AI tools (Gemini 2.0, YOLOv8) for precise and intelligent edits.
   • Outputs high-quality videos with minimal user intervention.

Installation and Setup

Prerequisites

• Python 3.8 or higher
• pip package manager
• Streamlit for UI
• Required libraries:
  • opencv-python
  • numpy
  • ultralytics
  • streamlit
  • google-generativeai
  • python-dotenv

Step 1: Install Dependencies

pip install -r requirements.txt

Step 2: Set Up Environment Variables

Create a .env file in the root directory and add your API keys:

GEMINI=<Your Gemini API Key>

Step 3: Run the Application

Start the Streamlit app:

streamlit run main4.py

Usage Instructions

1. Upload your video via the sidebar in the Streamlit app.
2. Enter your natural language prompt in the text box (e.g., "Enhance brightness when a person is detected").
3. Click Generate Content to process the video using AI-powered tools.
4. Download and preview the edited video.

Supported Functions

• Adjust Brightness: Modify the brightness within a specified time range.
• Enhance Contrast: Apply histogram equalization for better contrast.
• Sharpen Video: Increase frame sharpness using convolution filters.
• Trim Video: Remove specific time ranges.
• Slow Down/Speed Up Video: Adjust playback speed for selected segments.
• Remove Objects: Detect and remove objects using YOLOv8.

Code Structure

• video_editor.py: Core video processing functions.
• tool_executor.py: Maps AI-generated function calls to video processing methods.
• main4.py: Streamlit-based user interface for interacting with the system.

Example Prompts

• "Crop the parts of the video where you see birds."
• "Remove the potted plant from the background."
• "Enhance the brightness only when a person is detected."

Contributing

We welcome contributions! Please fork the repository and submit a pull request with your changes.

Acknowledgements

• The YOLOv8 model for object detection.
• OpenAI's Gemini API for natural language understanding.
• Streamlit for the interactive user interface.