There are many websites that make an animated version of our face, but they always do it from an image, not a video.
Now Artificial Intelligence can help us go one step further, and developers Tejas Mahajan and Niraj Pandkar have proven it.
They have created an artificial intelligence tool that can turn photos and videos into cartoons: Cartoonizer.
It is based on a research work carried out by researchers at the University of Tokyo. They use GPU (Graphics Processing Unit) servers for video inference and CPU (central processing units) for images, and this is what they can achieve:
Although not perfect, it could be used to create quick prototypes for anime, cartoons, and games. Logically the doors are open to create art, add short scenes to games and much more.
At the moment the result cannot be customized, and the videos must be less than 30 megabytes, but it can be tried for free and without the need for registration. On the same website there are other examples with videos of the Friends series, some photos and others.
Mahajan and Pandkar plan to open the source code and write an article on architecture in the next few days. If you are interested, stay tuned to their website and social media profiles.
