Skip to content

jberwald/IMA-Deep-Learning

 
 

Repository files navigation

IMA-Deep-Learning

deepsix

deepsix.collection

This subpackage contains modules for downloading images from various sources.

  • flickr (requires flickr API key)
  • imagenet

Each source has a function urls_tagged which returns a list of image URLs with a particular tag, keyword, synset id, etc.

Example usage:

human_urls = deepsix.collection.imagenet.urls_tagged('n07942152', max_images=5)
deepsix.collection.get_images_from_urls(human_urls, output_folder='downloads')

deepsix.anomalize

This subpackage contains modules for creating anomalies in an image, in order to provide a training set of "bad" images.

deepsix.preprocessing

This subpackage contains modules for loading images into nparrays and applying filters to examine various aspects of the image.

About

The IMA Workshop team working on neural networks for image anomaly detection

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • TeX 73.3%
  • Python 24.2%
  • Makefile 1.7%
  • Shell 0.8%