Japan Researchers Develop Automated Technique For Anime Colorization Using Deep Learning (brightsurf.com)

Japanese researchers have developed a technique for automatic colorization in anime production. "To promote efficiency and automation in anime production, the research team focused on the possibility of automating the colorization of trace images in the finishing process of anime production," reports Brightsurf. "By integrating the anime production technology and know-how of IMAGICA GROUP and OLM Digital with the machine learning, computer graphics and vision technology of NAIST, the research team succeeded in developing the world's first technique for automatic colorization of Japanese anime production." From the report: After the trace image cleaning in a pre-processing step, automatic colorization is performed according to the color script of the character using a deep learning-based image segmentation algorithm. The colorization result is refined in a post-process step using voting techniques for each closed region. This technique will be presented at SIGGRAPH ASIA 2018, an international conference on computer graphics and interactive techniques, to be held in Tokyo, Japan on Dec. 4-7. While this technique is still in the preliminary research stage, the research team will further improve its accuracy and validate it in production within the anime production studio. The product of this will be available for commercialization from 2020.

Comments

Re:How many frames

[mikael]

Disney offshored that work from Hollywood to the Far East decades ago, then brought it back when it was automated by computer. Artists from the art schools are way more qualified than to inbetween and fill in paint cells.

If you look at how they did many of those early animations, there were motion artists who took movies of dancers, rotoscoped the animation to get generic stick figures for reference. These frames were a closely guarded secret as they were reused between animations. When it came to make an animation, they took those reference frames, outlined each with a simple body, get detail artists do a fully detailed frame, then hand those over to inbetweeners who did the intermediate animation frames, and finally to the cel painters who outlined and filled in each individual frame. The deep learning system replaces these stages.

https://en.wikipedia.org/wiki/...