From: diego@minerva.st.dsi.unimi.it (Diego Montefusco)
Subject: TECH: Realistic Hand Model driven by DataGlove
Date: Tue, 29 Dec 92 23:04:11 CET



Hi all!

I've worked with a DataGlove for some time now (a base model: 10
sensors), and I'd like to know if anyone else who has worked with it
has come up with a good solution of a problem I found.

I'd like to have a good wire-frame model of the hand (but the same
problem should apply to solid modeled ones). By wire frame HERE I mean
that each finger is just a serie of 3 segments, not "strips".

The problem I found is the THUMB: since the DataGlove can't detect all
movements of the first joint of the thumb (i.e. the one which links
the finger and the hand), without additional sensors, I was wondering
around WHICH axis the thumb should twist about: it surely isn't one of
the coordinate ones (X,Y,Z), and I've come out with a strange one,
empirically (sp?) found with trial and error. But it still doesn't
satisfy me. Infact if I make an OK posture, touching thumb and index,
the on-screen hand doesn't do the same, having always some
mis-alignment (sp?) between the two fingers. It surely depends on the
fact that I LOCKED the thumb movement on the WRONG plane.

Has someone else found a better solution? Reply directly to me, and
I'll summarize if there's enough interest.

Thanks very much and HAPPY NEW YEAR to all the Net!

Diego

-----------------------------------------------------------------------------
 Diego Montefusco                        B E W A R E ! !
  Via Pirano, 4                        PLEASE REPLY ME TO:
  20127 Milano                    montefus@ghost.dsi.unimi.it
    ITALIA                            (not to the account
phone:  +39 (0)2 27001467             I'm e-mailing from!)
-----------------------------------------------------------------------------