What would Disney characters look like Anime style?
We, developers, have a lot of interesting when it comes to visual analysis nowadays. I was looking for a some fun project on Github and found this API. With this Open source, you can make selfie to anime by Machine Learning (AI).
The paper called 'UGATIT' and it was written by NCSOFT Vision AI Lab. 'UGATIT' means that Unsupervised Generative Attentional Networks with Adaptive Layer-Instance Normalization for Image-to-Image Translation.
As it said, unsupervised Image-to-Image translation, U-GAT-IT can generate undistorted image by focusing more on the distinct regions between source and target domain by exploiting the attention modules.
Thank you for reading!
and Enjoy it!
GitHub repository
http://bit.ly/2U2hO0B
API and website deployed on Ainize
http://bit.ly/390JkQr
Top comments (0)