How to Generate Waifu Art Using Machine Learning

This is a video developed by Everyone is Happy Productions which is a “boutique multidisciplinary design company providing mappings between existing and emerging technologies, arts workers and communities.” The video demonstrates the use of machine learning to develop lyrics and syncopated animation.

Waifu Synthesis Using Machine Learning

“All of the animation is made in real-time using a StyleGan neural network trained on the Danbooru2018 dataset, a large scale anime image database with 3.33m+ images annotated with 99.7m+ tags.”

A StyleGan (Style-Based Generator Architecture for GANs) is a machine-learning architecture which can be used to generate artificial imagery. A famous, and notorious example of a StyleGan is ThisPersonDoesNotExist.com which displays imagery of artificial faces produced by a computer.

We have seen how the use of deepfakes can be disruptive to the identification of faked news, however, the implications of this type of synthesized media in pop culture are equally as paradigm changing.

The development of this type of media should challenge us to ask important questions such as:

  • What are the potential opportunities for synthetic art?
  • What are the potential pitfalls of using synthetic performers in pop culture?

More info about the video is available here.
For more info about StyleGANs, check out this informative article.

 

Dennis Bonilla has been a user experience designer, software developer, and digital strategist. Dennis is the co-founder and Chief Technology Officer at Baltu Studios. Dennis co-founded Unified Pop Theory with his friends. Dennis is a trend finder and idea maker who is inspired by individuals that believe the world can be changed one great project at a time. Dennis can be reached on LinkedIn, Twitter, and Instagram.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Back to site top