If you’ve ever wondered what your animation would look like, you can now find out, as the developers Tejas Mahajan and Niraj Pandkar have created an AI-based tool to convert your photos and videos to animation.
The cartoonizer tool is based on a research paper prepared by two researchers from the University of Tokyo Xinrui Wang and Jinze Yu, where the artificial intelligence tool, Cartoonizer, takes advantage of its open source implementation to create a publicly available demonstration, and uses GPU servers , To infer video, and CPUs of images.
The artificial attempt of the Cartoonizer tool to turn people into animated is not accurate, but its videos are impressive. Check out this cartoon scene from Marvel’s Avengers:
Mahajan and Pandkar believed that the tool could create rapid prototypes for anime, animation and games, as they perceived to generate minimal art and add short scenes to games and would also assist animation designers in their work.
Mahajan and Pandkar plan to open source code and write an essay on architecture within the next few days.
The tool currently only handles up to 10 seconds of a video clip, and if you happen to upload a video larger than 10 seconds, only the first 10 seconds will be considered, the allowed video file size is 30MB, and the supported image file formats are: jpeg, png, the supported video file formats are: mp4, webm, avi, mkv g but images GIF / TIFF are not supported, and your data is deleted while the demo is running according to the tool.
The Cartoonizer tool can be used in some of the following areas:
- Produce fast animation and game models.
- Since the tool is generally subject to facial properties and information, it can be used to generate minimal art.
- Games can easily import brief scenes without using motion capture.
- The tool can be used as an assistant for animation designers.