art

Tweet Follow @compvision

Computer Vision Cinema is a project at the Politecnic University of Valencia, Spain, on "research using computer vision tools and techniques to generate new audiovisual languages." The website has a blog reporting the progress of the project, which uses the Kinect, OpenCV, OpenNI, and Nite. The video below shows an automated camera turret.

Related articles:

• OpenNI organization announced and natural interaction software released
• OpenNI launches developer Arena for gesture-based apps and games
• ZigFu provides rapid-setup motion capture

Tweet Follow @compvision

Dr. Enrico Costanza of the University of Southampton, U.K., has developed a marker-based music system. A computer vision system recognizes the types and positions of different markers to interactively create multiple tracks of music. The open-source software is available from d-touch.org. The system is demonstrated in the video below.

Wired Magazine published a detailed article on research by MIT professor Pawan Sinha on how the study of caricatures may lead to the development of better computer vision algorithms for face recognition. Since all faces have a similar structure, with relatively subtle differences, one theory of face recognition in the human visual system is that our brains identify the ways in which different people's faces depart from our internal representation of an average face. Experiments with human subjects suggest that caricatures can be more recognizable than unexaggerated representations. The theory would also explain the other-race effect, in which people tend to perceive people of other races than themselves as all looking alike, because people tend to form their perceptual average face from people of their own community. Professor Sinha's research is measuring which variables of departure from a mean face are the most significant for face recognition.

The J. Paul Getty Museum announced that it is the first museum to make its entire collection of paintings available via Google Goggles. With the Getty version of the Google Goggles app, visitors can point a smartphone camera at a painting to access information about that painting or its artist: much more information than a curator can place on a traditional wall plaque. According to the press release, the computer vision algorithms underlying the app are able to discriminate between different versions of the same painting.

Related article on Computer Vision Central:

• Mobile augmented reality enhances museum tour

Zachary Lieberman's exhibit, Making the Invisible Visible, at Georgia's Telfair museums, features several uses of imaging and computer vision technology for participatory performance art. One work by Lieberman and collaborator Golan Levin allows users to create sounds and shapes with their hands; a second uses a 3D scanner to create portraits of gallery visitors; and a third shows a video of an eye tracker that enabled people to write with their eyes.

New Scientist has an interview with Tod Machover, director of the Hyperinstruments and Opera of the Future groups at MIT's Media Lab. He has created the opera, Death and the Powers, in which one of the actors is backstage, monitored by computer vision, and his performance "is infused into the stage through lighting, images, sound and robotics. The set emotes." In this way, the set becomes active and interacts with on-stage characters.

Petronio A. Bendito, a professor in the Visual and Performing Arts at Purdue University, creates interactive art installations that use computer vision software to analyze the motion and appearance of visitor-participants. A recent installation called Action//Musique plays music according to the visitor's movements and determines colors in a wall projection according to the visitor's clothing. Additional information can be found in a JConline article.

Dancers in North American and Asia will dance together using computer vision software developed by Organic Motion.

Organic Motion's markerless motion capture technology will be used in a performance featuring dancers in Canada, Japan, and the United States. Each dancer's movements will be captured and animated at the same time, and synthesized into a single performance. Viewers can watch the original video and the animated version live, at 00:00 UTC/GMT November 4th, at CONNECTING the EDGES: An Inter-Continental LiveVibe. The performance is part of the International Digital Media and Arts Association (IDMAA) 2010 Conference in Vancouver, Canada.

Australian researchers from the University of Technology, Sydney, have developed dance routines for robots. The project involved fine-tuning motor control algorithms to enable fluid dance movements, and a new programming language to direct the routines. The NAO robots performed with the Melbourne dance troupe Wickid Force. More information and video are available in a Sydney Morning Herald article.

Submitted by Boaz Super

Syyn Labs, a company that builds installations combining art and technology, has demonstrated the Cloud Mirror at Sundance. It's an augmented reality installation that combines a person's mirror image with photos and text from that person's online existence in sites such as Twitter and Facebook. Computer vision identifies a distinctive badge worn by each user and composites the photos and text in real time with the user's image. A video of Cloud Mirror can be viewed at the link above.