Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/ilirosmanaj/detect_kermit/llms.txt

Use this file to discover all available pages before exploring further.

download_from_google.py uses the google_images_download library to fetch extra images of Kermit from Google Images, supplementing the frames extracted from video with a broader variety of poses, lighting conditions, and contexts. This diversity helps the model generalise beyond the specific visual style of The Muppets episodes used for training.

​
Source code

helpers/download_from_google.py
import google_images_download
import ssl

from google_images_download import google_images_download


def main():
    ssl._create_default_https_context = ssl._create_unverified_context
    response = google_images_download.googleimagesdownload()

    # download by keyword
    arguments = {'keywords': 'kermit', 'limit': 100, 'output_directory': 'data/google-images'}
    paths = response.download(arguments)
    print(paths)


if __name__ == '__main__':
    main()

​
How to use

cd helpers
python downloads_from_google.py
Images are downloaded automatically and organised into data/google-images/kermit/ — the library creates a subdirectory named after the keyword.

​
Configuration

The arguments dictionary controls what gets downloaded:
KeyDefaultDescription
keywords'kermit'The search query sent to Google Images.
limit100Maximum number of images to download in a single run.
output_directory'data/google-images'Root directory where images are saved. A subdirectory named after the keyword is created automatically inside it.
Edit any of these values directly in the script before running to customise the search or change the output location.
The line ssl._create_default_https_context = ssl._create_unverified_context disables SSL certificate verification globally for the duration of the script. This is a workaround for environments where certificate validation fails when reaching Google’s servers. It is only appropriate for local development and data-collection workflows — do not carry this pattern into production code.
google_images_download is listed in requirements.txt and is installed automatically along with all other project dependencies when you run:
pip install -r requirements.txt

​
After downloading

The downloaded images are raw search results and may include irrelevant or low-quality content. Before using them for training you should:
  1. Manually review the images in data/google-images/kermit/ and remove anything that doesn’t clearly show Kermit.
  2. Move the accepted images into the appropriate training subdirectory: 
    mv data/google-images/kermit/*.jpg data/images/train/kermit/
  3. Ensure the final folder structure matches what ImageAI expects before starting a training run.

Build docs developers (and LLMs) love

Get started for freeTalk to us