home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!europa.asd.contel.com!emory!wupost!spool.mu.edu!uwm.edu!ogicse!news.u.washington.edu!stein.u.washington.edu!hlab
- From: rnm@well.sf.ca.us (Robert Marsanyi)
- Newsgroups: sci.virtual-worlds
- Subject: Re: APPS: Gesture recognition using simple statistics (DataGlove app.)
- Message-ID: <1992Dec13.030320.13954@u.washington.edu>
- Date: 13 Dec 92 00:13:28 GMT
- Article-I.D.: u.1992Dec13.030320.13954
- References: <1992Dec2.082956.24747@u.washington.edu> <1992Dec5.095830.14950@u.wa
- Sender: news@u.washington.edu (USENET News System)
- Organization: Whole Earth 'Lectronic Link
- Lines: 18
- Approved: cyberoid@milton.u.washington.edu
- Originator: hlab@stein.u.washington.edu
-
-
-
- > ..discussion of static gesture recognition and extensions for including
- > dynamics..
-
- Note also Dean Rubine's work at Carnegie-Mellon on dynamic single-path
- gesture recognition. He sent me a copy of his thesis on request, and it
- seems, although written for 2-D (mouse) input, to be perfectly extensible to
- 3-D input. The recognizer is fast, too (based on timings in the document),
- and the training process is again training-by-example, with a small example
- set providing good recognition.
-
- I'm set to hack on this stuff early next year, when my schedule clears out.
-
- But if somebody beats me to it ...
-
- --rbt
- (Robert Marsanyi, rnm@well.sf.ca)
-