rkp64 / videointerpreterai

Uses GPT-4 vision model to interpret videos

Video Understanding App

This Streamlit-based application integrates computer vision and AI to analyze and understand video content. It uses OpenAI's powerful GPT-4 API to generate descriptive text for video sequences, offering a glimpse into the content and context of the videos.

Features

• Video Upload: Users can upload videos in MP4, AVI, or MOV formats.
• Frame Display: The application displays the first frame of the uploaded video.
• Automated Description: Leverages OpenAI's GPT-4 to generate descriptions for the video content.

How to Use

1. Upload a video file using the Streamlit file uploader.
2. Wait as the app processes the video and displays the first frame.
3. Read the generated description of the video content.

Technical Details

• The app saves the video to a temporary file for processing.
• Frames are extracted and converted to base64 for analysis.
• Descriptions are generated by sending frames to the OpenAI API.
• The first frame of the video and its description are displayed on the UI.

Getting Started

To run this application locally, follow these steps:

1. Ensure you have Python installed on your system.

2. Clone this repository to your local machine.

3. Install the required dependencies

4. Create a .env file in the root directory of the project and add your OpenAI API key:

   OPENAI_API_KEY='your_api_key_here'
   

5. Start the Streamlit app:

   streamlit run your_script_name.py

Replace your_script_name.py with the name of your Python script.

Installation

Make sure you have the following Python packages installed:

• streamlit
• opencv-python-headless
• base64
• tempfile
• openai
• dotenv
• requests

You can install these packages using pip