What is it about?

Anime is a hand-drawn animation style that originated in Japan. The style is unique and attracting worldwide attention. The purpose of this research is to create a transformation formula to stylize Computer Graphics (CG) motions as anime. The proposed method analyzes the motion characteristics from an input motion capture data and automatically converts it into a motion suitable for lower frame rates. The key to the conversion lies in the omission of the redundancy of speed changes inherent in real motions.

Featured Image

Why is it important?

Anime’s unique style of motion has influenced 3DCG animations. For example, non-photorealistic CG animation attempts to mimic the style of anime motions by reducing the number of frames per second (fps), in addition to pursuing aesthetic. However, because joint motions are complex, they are sometimes created from physically captured motions. Further, because these motions have a small speed strength, lowering the fps constantly does not result in an anime-like motion, but rather reduces the quality of the motion, where the jerkiness of the motion causes discomfort. To make the motion more attractive with fewer drawings, the timing, spacing, and camera shot angles must be adjusted according to the unique motion drawing techniques established by Japanese animators over many years.

Perspectives

Compared to high frame rate animations, anime is drawn with the consideration that a single picture is displayed for a longer time, making momentary poses more impressive. In non-photorealistic CG animations, I hope that the distribution of frames will be ambitiously recognized as an essential choice to bring out the individuality of artistic works.

Yui Koroku
Keio University

The skills of master craftsmen can be algorithmically represented, actively preserved, and shared. This is also true for anime-like motion synthesis.

Prof., Dr. Issei Fujishiro
Keio University

Read the Original

This page is a summary of: Anime-Like Motion Transfer with Optimal Viewpoints, December 2022, ACM (Association for Computing Machinery),
DOI: 10.1145/3550082.3564212.
You can read the full text:

Read

Contributors

The following have contributed to this page