Spectrum, exploring the gender continuum with deep learning
Clone or download
Latest commit cea11b1 Jan 21, 2018

README.md

Spectrum

exploring the gender continuum with deep learning

Videos are posted to the eXs youtube channel.

Notes for a new video

  • Download the video from the url youtube-dl URL
  • Convert the images, one per second mkdir out; avconv -r 4 -an -y "out/%04d.png" -i VIDEO.webm
  • Manually remove frames at the start and end that are not content
  • Identify the faces from the video python segment_faces.py videos/drag_queen/out/.
  • Use the CNN model to label each image python predict_images.py image_processed/videos/drag_queen/out/. Data is saved in the results folder.
  • Generate the animation python render_animations.py drag_queen
  • Update the video sources in this README.
  • Upload to youtube and fill out the sources template.

Attribution

Tools and assests for the Spectrum project. Model was trained off images from IMDB-WIKI dataset. Code was adapted from BoyuanJiang from the the paper, Deep Expectation of Real and Apparent Age from a Single Image Without Facial Landmarks.

Video sources

Drag Queens

Drag Kings

Control

Install notes