Watch What You Are
About
Watch What You Are is an immersive computer vision system that links the 2-D virtual world of movies and TV to the 3-D reality of the user. The game starts when the vision tracking system looks for pattern matches between the body position of a movie character and a real time video feed of the participant. When the participant enters into the camera's field of focus, the beginning of the film (i.e. “Superninjas”, a 70's kung fu movie) is triggered. The main movie sequence continues to play until a break point occurs where the participant is challenged to mimic the movie character's position.
More info on Justin's site.
Team
Justin Downs and Rucyl Mills (New York) study at NYU's Interactive Telecommunications Program.
Collaborators:
eric mika
dan ribaudo
subalekha udayasankar
jessica banks
Photos
Next
Friday, July 4
- reprogrammed application in openFrameworks with Zach
- structured flow of video for each channel
- tested image capturing with floor covering
Monday, June 30
- camera n lighting
- Video - MAX/MSP
- feedback - right n' wrong pixels.. (other than using the grid)
audio feedback ( music while the person is interacting)
Process
- secure and old television (dial vs. numpad vs remote?)
- set up test prototype to sort out IR and allow basic user testing (basic camera + code + projector feedback loop)
- refine and optimize code, possibly port image processing code from Java to C++. improve visual interface, e.g. how do you know when you're in / getting close to position – reduce frustration.
- find / capture / curate several channels of video
- cardboard / foamcore props
- …
Video Sources
- YouTube Hi-Res Hack: add &fmt=18 to the end of the url Lifehacker Post
