I found this project a minute ago and thought it might be interesting to those of you who haven't seen it yet. The panel is a matrix of infrared LEDs and a hand is placed between the panel and a camera, the camera records the remaining unobscured LEDs and I think a computer takes over, analyzing the images from the camera. This is a novel and simple way to detect hand signs... but I don't think the approach would be very useful in finger chording. Instead I would say more intense image processing of color video from a webcam is required. Finger chording recognition ultimately boils down to using images to extrapolate information about finger positions relative to other fingers. The ratios of relative finger positions found in the analysis of these images is the basis for the information considered in determining the actual chord being signed. So, figuring out how the heck you measure a fingers position is paramount. One thing is for sure, this all goes back to computer vision stuff. If anyone has any referrences for computer vision that would be helpful to this thread. There is a list of other projects at the Contextual Computing Group's site = here: http://www.cc.gatech.edu/ccg/research.html _______________________________________________ Wear-Hard mailing listhttp://www.haven.org/mailman/listinfo/wear-hard
From Wear-Hard Mailing list Archive (WH)
Maintained by R. Paul McCarty
Archive created with babymail