Publication date: 
2017/10/30
A prestigious Adobe MAX 2017 conference was held in Las Vegas from 18 to 20 October. , The Puppetron project, created in cooperation with the Department of Computer Graphics and Interaction of the Faculty of Electrical Engineering, CTU and Adobe, was presented there. The revolutionary algorithm allows you to transfer an art style from hand-drawn artwork to portrait photography to faithfully preserve both identity of the stylized person and basic characteristics of the art media used. This is how the platform differs from applications based on the use of neural networks such as Prisma or DeepArt.

The software, created by the team of doc. Daniel Sýkora, can faithfully translate the art style from an existing hand-painted portrait into a real human face video. The technique can preserve the identity of the subject as well as the visual richness of model painting. Additionally, it allows you to control the visual dynamics of the image over time, and thus bring it closer to the look of otherwise very laborious total animation where each frame is drawn by hand. This is the first successful transfer of technology developed at the Faculty of Electrical Engineering at the Czech Technical University to an Adobe professional tool.

A video from Puppetron presentaion at the Adobe Max conference is available here.

Picture

Picture

Picture