An AI-powered tool that scrapes images from a webpage and generates captions locally for each image using the BLIP model. This project uses Streamlit to create an interactive web interface.

The main homepage where users can input a URL to scrape images from.

The downloaded image links with generated captions.

Objective

Captionator is designed to help users automatically generate descriptive captions on their local machines for images found on a webpage. This tool can be useful in the following contexts:

1. Content Creation: Content creators, bloggers, or social media managers can use this tool to quickly generate captions for images on their websites or blogs. It saves time and adds valuable metadata to images.
2. Accessibility: The captions generated can be used to provide alt text for images, helping visually impaired users understand the content on a webpage.
3. Image Dataset Creation: Researchers and data scientists can leverage this tool to build datasets with image captions, which can be useful for training machine learning models in computer vision tasks.
4. Web Scraping & Automation: This tool automates the process of scraping images and generating captions, which can be useful for businesses or organizations that need to collect large amounts of image data from various websites.

How It Works

1. Input: Paste the URL of any webpage containing images.
2. Processing: The app scrapes all images from the page, processes them using the BLIP (Bootstrapping Language Image Pretraining) model, and generates captions.
3. Output: You can download the generated captions as a .csv file containing the image URLs and their respective captions.

Features

• Scrapes all images from the given URL.
• Generates captions for each image using an AI model (BLIP).
• Download captions in .csv format for further use.
• Interactive web interface powered by Streamlit.

Setup and Installation

1. Clone the repository:

   git clone https://github.com/bushraqurban/Captionator.git
   cd captionator

2. Create a virtual environment and activate it (optional but recommended):

    # On Mac/Linux:
    python3 -m venv venv
    source venv/bin/activate

   # On Windows
   python -m venv venv 
   .\venv\Scripts\activate

3. Install the required dependencies:

   pip install -r requirements.txt

4. Run the application:

   python3 app.py # On Mac
   python app.py # On Windows

   This will launch the app in your browser.

How to Use

1. Paste any webpage URL that contains images (e.g., Wikipedia or blogs).
2. Click the Generate Captions button to generate captions.
3. After the captions are generated, download the captions file with .csv formate.

Example Output

Here’s an example of how the .csv file will look like after running the tool:

License

This project is licensed under the MIT License - see the LICENSE file for details.

Project Acknowledgment and Enhancements

This project was inspired by the IBM AI Developer Professional Certificate course guided project. I have further enhanced it by adding several custom features, including:

• A user-friendly interface that allows users to interact with the app directly without needing to run Python scripts.
• An improved output format that generates captions in a CSV file with a table structure, making it more organized and user-friendly.

Technologies Used

• Python
• BLIP Model
• Streamlit

---