Dazzling new techniques in computer graphics and provocative new ideas about interaction were shown off by researchers from academia and industry at the Siggraph conference in Anaheim, California, last week. Face Replacer
New software tracks a person’s facial expressions and maps it—in real-time—onto a digital character, whether that might be an alien or an ape. The technology could make it easier to create computer animated characters that closely match the expressions of an actor.
To get started, just shake your head once in front of the Kinect sensor the software uses so it can capture an impression of the shape of your face. The mapping gets better over time, because as a person moves his or her face, the system works to learn its shape better, producing a more accurate mimic. The software was created by Hao Li, Jihun Yu, Yuting Ye, and Chris Bregler of Industrial Light and Magic.
See a video of the face-replacing software. Interactive Water
The surface of tinted water becomes an interactive display using a system called AquaTop, created by Yasushi Matoba at the University of Electro-Communications in Tokyo. A projector creates the imagery, which can be controlled by hands or fingers tracked using a Kinect depth sensor. AquaTop can also detect fingers protruding from below the water’s surface, and position imagery around them. Water Games
Another view of AquaTop shows a two-player game where a virtual ball is controlled by swatting at it on the water’s surface. AquaTop’s creator, Matoba, was inspired by the way bath salts that color water, popular in Japan, seem to create a blank canvas around a bathing person.
See a video of the AquaTop display. Air-Blast Feedback
This device fires puffs of air at a gamer’s hands or body to provide physical feedback when he or she is playing a gesture-controlled game. The device, called Aireal, can pivot to blast its target, and fires dense rings of spinning air that can travel large distances without breaking up. The system was designed at Disney’s research labs in Pittsburgh, and the system’s creators believe the air-blast approach could be used for gaming, immersive storytelling, and even mobile interfaces.
See a video of the Aireal device.
Software developed at Dartmouth College and UC Berkeley detects fake photos by checking that shadows in an image are consistent with the same light source. This image of Buzz Aldrin standing on the moon in 1969 passed the test, of course.
Research has shown that humans are poor at detecting inconsistency in shadows, suggesting that forgers probably don’t get them right much of the time. Eric Kee and Hany Farid of Dartmouth College created the software with James O’Brien of the University of California, Berkeley. They say software like theirs could help restore trust in photos in an era of easy digital manipulation.
Read more about the fake-photo detecting software.
Although 3-D printing is becoming cheaper, designing printable objects hasn’t gotten much easier. Software called Make It Stand aims to simplify the process for designers by ensuring their models will balance, no matter how complex they are. The software makes changes to a model’s volume so that it can balance perfectly on the tiniest of supports, while preserving surface details. Make It Stand was created by Romain Prévost, Emily Whiting, and Olga Sorkine-Hornung of the Swiss Federal Institute of Technology Zurich, and Sylvain Lefebvre of the National Institute for Research in Computer Science and Control, France.
See a video explaining the software.
Hair is one of the most challenging things for computer animators to simulate realistically. Software that works out the 3-D structure of the strands making up a hairstyle from a collection of photos could give future computer-generated characters more realistic coiffures.
This image illustrates how a series of 30 photos of the hairstyle on the left was converted into a detailed 3-D model using new software created by Linjie Luo and Szymon Rusinkiewicz of Princeton University, with Hao Li of Industrial Light and Magic. First the software works out a likely collection of 3-D strands that could make up the hairstyle (center) and how they would move, then creates a detailed simulation (right) that moves realistically.
See an image of the hair capture software. TV-Busting Games
A project called IllumiRoom enables a console game to escape the bounds of a TV screen, overlaying imagery onto the space around it. The system was developed by researchers at Microsoft and uses a Kinect depth sensor and projector positioned in front of a gamer to detect the shape of the room around the TV set, and overlay imagery onto it.
In tests so far, IllumiRoom has been used to add a new atmosphere to games by, for example, creating the illusion of falling snow around a TV, and having explosions extend beyond the screen.
See a video of IllumiRoom in action.