[PD] Kinect mappings

[Matthias Kronlachner]

hi!

if you have installed latest unstables of openni and nite it should  
start tracking of a new user almost immediately without the need of a  
calibration pose.
what takes long time (at least on my machine) is the initialization  
process. (when you start gemwin)
for me pix_openni works much better and stable on linux.

if you have gem framerate at 30fps you get 15x30 joints with x,y,z  
coordinates per second.
it's really not an easy task to extract useful information out of it.
i lately added the skeleton smoothing feature, that can be very  
helpful to smoothen the output. (skeleton_smoothing 0.0-1.0)

what you need is a set of abstractions to make geometric calculations  
from the datastream, the ability to monitor/draw graphs and the  
"interpretation" of this data/gesture detection. iemmatrix can be  
practical for the vector calculations and to make the patches more  
readable.
i'm working in the moment on such abstractions but still a work in progress.



anyway i'm curious how the kinect for windows sdk is working.
with the built in speech recognition and source localization over the  
microphone array it has some advantages over openni/nite.
but for me the platform dependency is a no-go. so i won't work on an  
external with this framework in the future.
maybe someone else? :-)


matthias


Zitat von "Pagano, Patrick" <pat at digitalworlds.ufl.edu>:

> I find that openni on the Mac takes to long to register the skeleton  
> so we are using processing to track and osc the joints and  
> relationships. I was originally just soliciting fun ideas for the  
> huge data stream the kinect provides
>
>
> Patrick Pagano B.S.,M.F.A
> Digital Media Engineer
> UF Digital Worlds Institute
> (352)294-2020
>
>
> On Mar 18, 2012, at 5:49 AM, "philippe boisnard"  
> <philemon1 at mac.com<mailto:philemon1 at mac.com>> wrote:
>
> Yes
>
> I use pix_openni, it's for me in fact, the better solution.
>
> p
>
>
> Le 18 mars 2012 à 07:02, dreamer a écrit :
>
> Why not use pix_openni to do the skeleton-tracking needed?
> Then you can omit OSCeleton, which is quite limited anyway.
>
>
>
> On Sun, Mar 18, 2012 at 3:52 AM, Pagano, Patrick  
> <pat at digitalworlds.ufl.edu<mailto:pat at digitalworlds.ufl.edu>> wrote:
> Hans
> I think i looked at this before. Do you have DIlib? It seems like  
> it's not available yet?
> I thought i grabbed this before but that was on linux
>
> pp
> ________________________________
> From: Hans-Christoph Steiner [hans at at.or.at<mailto:hans at at.or.at>]
> Sent: Saturday, March 17, 2012 10:12 PM
> To: Pagano, Patrick
> Cc: pd-list at iem.at<mailto:pd-list at iem.at>
> Subject: Re: [PD] Kinect mappings
>
>
> Check out William Brent's dilib:
> http://williambrent.conflations.com/pages/research.html
>
> .hc
>
> On Mar 13, 2012, at 1:23 PM, Pagano, Patrick wrote:
>
> Hello
>
> I am collaborating with a programmer who has developed a kinect  
> streamer with windows SDK that sends data over udp.
> I am soliciting some ideas on WHAT to do with the data stream.
> We are thinking movies/models and midi, but I would love to hear  
> some ideas regarding what to do with the data
>
> Cheers~
>
> pp
>
> <OSCroute_kinect4.pd>_______________________________________________
> Pd-list at iem.at<mailto:Pd-list at iem.at> mailing list
> UNSUBSCRIBE and account-management ->  
> http://lists.puredata.info/listinfo/pd-list
>
>
>
> ----------------------------------------------------------------------------
>
> The arc of history bends towards justice.     - Dr. Martin Luther King, Jr.
>
>
>
> _______________________________________________
> Pd-list at iem.at<mailto:Pd-list at iem.at> mailing list
> UNSUBSCRIBE and account-management ->  
> http://lists.puredata.info/listinfo/pd-list
>
>
> _______________________________________________
> Pd-list at iem.at<mailto:Pd-list at iem.at> mailing list
> UNSUBSCRIBE and account-management ->  
> http://lists.puredata.info/listinfo/pd-list