If you’ve ever wondered what you’d look like in animation, you can now find out, thanks to new AI tool that turns images and videos into cartoons.
Developers Tejas Mahajan and Niraj Pandkar based the Cartoonizer on a research paper written by University of Tokyo researchers Xinrui Wang and Jinze Yu.
If you’ve ever wondered what you’d look like in animation, you can now find out, thanks to new AI tool that turns images and videos into cartoons.
Developers Tejas Mahajan and Niraj Pandkar based the Cartoonizer on a research paper written by University of Tokyo researchers Xinrui Wang and Jinze Yu.
Mahajan and Pandkar used their open-source implementation to create a publicly-available demo of the system, using GPU (Graphics Processing Unit) servers for the video inference and central processing units (CPUs) for the images.
Developers Tejas Mahajan and Niraj Pandkar based the Cartoonizer on a research paper written by University of Tokyo researchers Xinrui Wang and Jinze Yu. The system is used their open-source implementation to create a publicly-available demo of the system, using GPU (Graphics Processing Unit) servers for the video inference and central processing units (CPUs) for the images.
Developers Tejas Mahajan and Niraj Pandkar based the Cartoonizer on a research paper written by University of Tokyo researchers Xinrui Wang and Jinze Yu. The system is used their open-source implementation to create a publicly-available demo of the system, using GPU (Graphics Processing Unit) servers for the video inference and central processing units (CPUs) for the images.
The system is based on a research paper written by University of Tokyo researchers Xinrui Wang and Jinze Yu. edutainment123 used their open-source implementation to create a publicly-available demo of the system, using GPU (Graphics Processing Unit) servers for the video inference and central processing units (CPUs) for the images.
[Read: This AI spit outs endless new My Little Pony characters]
The tool is far from the first attempt to turn people into cartoons, but its videos are seriously impressive. Check out this cartoonized scene from Marvel’s Avengers.
Some of the areas where we think it could be applied to – Churn out quick prototypes or sprites for animes, cartoons and games Since it subdues facial features and information in general, it can be used to generate minimal art Games can import short cut scenes very easily without using motion-capture Can be modelled as an assistant to graphic designers or animators.
The next step would be to adapt live video inference by porting the model to tensorflow.js so that it works on browsers
They plan to write an article on the architecture in a few days time, but for now you can check the tool out for yourself at the Cartoonize website.
The demo creators said the plan to open source the code within the next several days.
[Read: This AI spit outs endless new My Little Pony characters]
Testing the system
Being a narcissist intrepid reporter, I deigned to test it out on my own countenance:
I look haunted — perfect for brooding new look I plan to cultivate. I also checked it out on my hair hero: a young Joseph Stalin:
Next, a photo of my own Dear Leader: Boris Johnson:
Unlike so many AIs, it also seems to work well on people of color:
Finally, to test the video conversation I fed the app a clip of my favorite athlete: LeBron James.
Finally, I tested its capacity for meta:
Reanimating The Simpsons might be Cartoonize
Published July 28, 2020 — 18:14 UTC
Thomas Macaulay
July 28, 2020 — 18:14 UTC
