Automatically identify distinct scenes in a video and find 'visually similar' shots using an image classifier trained on Wikipedia data. See result:
Shell Python
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.


A Python script that automatically identifies distinct scenes in a video and generates video clips and thumbnails for each scene.

In addition to chopping up a video into scenes, you can also find 'visually similar' images from an image classifier trained on Wikipedia.

The classifier was built and trained by Douweo Singa. Find it here:



  • Set up a virtual environment & install requisite Python packages.
  • Put video file (format: mp4) in project folder.


  • Activate virtual environment. Make sure dependencies are properly installed using pip.
  • Run "" script to identify distinct scenes and generate video files.
  • Run "" script to generate a screen grab for each scene (format: jpg).
  • Run "" to download images that are 'visually similar' to the scene thumbnails.