On July 29, AI venture Radius Five (Shinjuku-ku, Tokyo) launched “Anime Art Painter,” a service that uses AI to generate art backgrounds for anime from photographs. You can convert it to an anime-like set just by uploading a photo. Generate four types of scenes from one image in about 30 seconds. Background generation is free, but image downloads are charged and start at 480 yen.
Roughly, deep learning is used to simplify the texture of photos. It is said to be a mechanism for converting to a style used in anime and illustrations. It is said that the texture conversion and color tone will be adjusted according to the objects such as buildings, night views, natural objects such as mountains, objects taken at a short distance, and indoor backgrounds.
The four types of images generated are two types of watercolor-style paintings used in animation and two types of animation-style pictures close to the solid painting.
The company points out that the current anime industry is “stuck in production, which is said to have been waiting for two years.” Generally, anime uses 200 to 300 art backgrounds per episode, but the industry lacks human resources to draw the background. The company explained the development process, saying, “I hope this AI will help improve the man-hours for animation production.”
So far, the company has provided “AI that converts photos into painting-like illustrations of Van Gogh and Picasso,” “AI that converts facial photos into caricatures in 30 seconds”, and “services that can convert animation to high resolution.”