NVIDIA StyleGAN AI Used to Create Tezuka-like Characters in New ‘PHAEDO’ Manga


StyleGAN, the generative adversarial network created by NVIDIA and open sourced in February 2019, was used to create characters inspired by the works of the father of manga, Osamu Tezuka, in the new manga PHAEDO (released in the weekly comics magazine Morning).

Project leader Ryohei Orihara, senior fellow at the Digital Process Innovation Center at Kioxia, stated:

AMD Reaffirms 30x Energy Efficiency Goal For Data Centers By 2025

Future Memories 01, ‘Tezuka 2020’, is a project in which AI and humans create a piece of work from Tezuka’s many memories. Manga has become well-known as a manifestation of Japanese culture around the world, and it is read and loved across generations.

Satoshi Kurihara, professor at the Faculty of Science and Technology at Keio University and professor of AI at the University of Electro-Communications, added:

In order to have AI learn the characteristics of Osamu Tezuka, we used manpower and AI that specializes in image recognition to convert the parts of characters’ faces into data and analyze the development of various scenarios. With AI learning the characteristics of Osamu Tezuka through this data, we generated characters and storylines that have Tezuka’s characteristics.

Before using NVIDIA StyleGAN, a training set had to be compiled. Professor Kazushi Mukaiyama from Future University Hakodate compiled over ten thousand facial images from Tezuka's esteemed works. Mukaiyama said:

I had the computer learn character faces outside of Tezuka’s works as well as actual human faces, and I made it learn the augmented data generated by flipping the characters’ faces left and right.

Training and inference were done through multiple NVIDIA V100 GPUs and, of course, the cuDNN-accelerated TensorFlow deep learning farmework.

The results of this 'transfer learning' are surprisingly good, as you can see for yourselves in the image and video below. If you want to learn even more about this ambitious creative process, check out the website dedicated to the ‘Tezuka 2020’ project.