Manga generator creates comics from your posture
If you've longed to be the star of a manga comic, here's your chance to punch, jump and kick your way into one. Shirai Lab's Manga Generator turns any pose you strike into a comic panel in real time with the help of a Kinect sensor, adding backgrounds, props, speech bubbles and sound effects to convey the emotion you want to express. After a couple of minutes you're given a single-page multi-panel comic that features your image, fancifully rendered manga-style throughout.
Developed by Akihiko Shirai and his team at Kanagawa Institute of Technology, Japan, the technology lets players select a story and take on the role of a character in it. The manga frame, story background and assorted comic effects are projected onto an 80-inch screen, giving players a full view of what the comic panel looks like.
The Kinect sensor tracks the player's movements and they get to see themselves move around in the comic panel, kind of like the moving photographs in the Harry Potter books. The real-time "shader" software draws them as they move within the scene, and when they strike a pose they like and hold it, it's captured and rendered into the comic panel.
The whole effect is pretty neat, as you get to see a moving drawing of yourself within a scene, giving you the option to try out various poses within a set time. The software gauges your emotion from your posture and arranges the comic effects accordingly. It notices how you bend your elbows, hips and shoulders to figure out the kind of emotion you might be feeling.
For instance, pointing may indicate puzzlement, so a question mark and other visual effects may be rendered. Jumping up may reflect success, so a "yahoo" speech bubble and a sunburst pattern might be drawn in. Once a panel is completed, the same process is repeated for every panel until the story is completed and printed out. Since the comic effects change according to the player's posture, every personalized manga comic is unique.
"The biggest feature of this system is, it observes a person’s posture, chooses a background to match, and synthesizes the two,” says Shirai in a Diginfo video. The developers plan to improve the shader software to deliver better quality images. They're also excited about the progress they've made with getting the layout software to arrange speech bubbles and props at appropriate locations.
They hope the technology will find use in interactive e-books, stories and digital signage and also attract advertisers who might want to print their ads on the comic's reverse side. The Manga Generator was presented at the Laval Virtual conference earlier this year.
Check out Diginfo's video of Shira Lab's Manga Generator below