From: Craig Hubley Subject: real product (was Re: Video input for (hand) motion sensing.) Date: Thu, 17 Oct 1991 18:19:48 -0400 Organization: UTCS Public Access In article <1991Oct16.145820.26485@milton.u.washington.edu> eeyore@dcs.qmw.ac. uk (Mark Anthony Brown;E200) writes: >In <1991Oct12.202324.15863@watserv1.waterloo.edu> mksd423f@shine20.uwaterloo. >ca (SD423F Course - Sys Design) writes: > >>What about some sort of video interface as an INPUT to a VR system? By >>pointing a video camera at a given area, there could be ways of interpreting >>the image to infer what the user is doing. It may involve multiple cameras, >>distance sensors, or chromakey techniques, but the main idea would be to >>avoid using anything which would actually be in contact with the person. All of these have been used. >Video input is used almost exclusively in Myron Kreuger's artificial reality >systems like VIDEOPLACE. His video perception system is described in his >book "Artificial Reality II" (Addison-Wesley, 1991). Kreuger's system is the most mature but also impossible to implement without all his special hardware. He is really stuck on his old wire-wrap boards for some reason... >Krueger's system uses a single video camera to extract the user's silhouette, >which is combined with a model of the human body and the previous silhouettes >to locate the extremities (the past silhouettes are required to help to >eliminate the ambiguity of the image). However, multiple video cameras could >be used to locate the extremities unambiguously. Yes, and I would be interested in finding any references to someone doing this to create a 3D array of positional information. If nothing else it would be good for redundancy with dataglove/suit information. An ideal system would not care from which devices it was getting its position info. In the meantime if you want to buy a system today that allows you to do this kind of 2D video input on a CHEAP PLATFORM (Amigas and I think soon Macs), contact Vivid Effects (catinhat@well.sf.ca.us but physically in Toronto) who have had this working since 1987. They were selling it at CHI'91 earlier this year. There are lots of examples of their work in galleries, etc. - one at Tour of the Universe in Toronto and another at the Smithsonian, and lots more they could tell you about. As others have pointed out, there are lots of excellent tools for doing video-based stuff on Amigas. David Rokeby's Very Nervous System also uses video input but with less of an emphasis on mirroring exact positions of extremities and more on catching parameters that describe the entire motion, since he's converting the input directly to music. There is also Kevin Landell and someone at CMU who built a "videoharp" but I don't have any exact pointers to these guys. Hope this helps, Craig -- Craig Hubley -- Consultants in object-oriented technology & techniques, -- Craig Hubley & Associates -- user interface design & user productivity -- craig@gpu.utcs.Utoronto.CA UUNET!utai!utgpu!craig craig@utorgpu.BITNET craig@gpu.utcs.toronto.EDU {allegra,bnr-vpa,decvax}!utcsri!utgpu!craig 28 First Avenue, Toronto, Ontario M4M 1W8 Canada (US TC) call 416-466-4097