[PD] Body tracking

Tue Jan 11 23:34:39 CET 2011

Hello all,

I'm new to this list... I've just started out with PD, with body-tracking in
mind actually.

I'm working with the kinect - you can get it working on your OS of choice,
via the wonderful community effort at:
http://openkinect.org/wiki/Main_Page

Seems like a natural fit for PD, and people are actually connecting it to
the OpenNI project here: http://www.openni.org/

and there are some other skeleton tracking efforts that have been discussed
on the openkinect google group.

Hope this helps!

Best,
Ali

On Mon, Jan 10, 2011 at 6:06 AM, Jaime Oliver <jaime.oliver2 at gmail.com>wrote:

> > the truth is: it depends on the purpose.
>
> Nothing Truer...
>
> >> Not defending anything, but just to remind 60 fps of a pointgrey camera
> is not slow.
> >for gesture, i consider everything under 200Hz as slow.
>
> surely this is the case for percussion as in your example here:
>
> > by example, things like this have to be done at 1000Hz:
> > http://vimeo.com/9598075
>
> However, with a minimum of 75fps with a ps3 camera (you can get up to
> 120fps at 320x240...) you can get very responsive results:
> www.jaimeoliver.pe/sc2
>
> It is clear that this project is not as concerned with precise triggers, it
> is more concerned with continuous control...
>
> you just need to find out if what these space&temporal resolutions offer
> you what you're looking for. (+ solving less robust lighting, which is what
> really determines if video tracking is the way to go...)
>
> best,
>
> J
>
>
>
>
>>
>>
>>
>> On Sun, Jan 9, 2011 at 12:37 AM, cyrille henry <ch at chnry.net> wrote:
>>
>>>
>>>
>>> Le 08/01/2011 20:24, Pedro Lopes a écrit :
>>>
>>>  Not defending anything, but just to remind 60 fps of a pointgrey camera
>>>> is not slow.
>>>>
>>> for gesture, i consider everything under 200Hz as slow.
>>> including the wiimote 100Hz.
>>>
>>> by example, things like this have to be done at 1000Hz:
>>> http://vimeo.com/9598075
>>> (we tried with slower system but it was not possible to be accurate
>>> enough).
>>>
>>>
>>>
>>>  you probably will have a lot of trouble in dealing with 60 frames trying
>>>> to cycle through your code every second.
>>>>
>>> you have problem analysing the image, but dealing with the data is no
>>> problem.
>>>
>>>
>>>  Accels and other hardware sensors are very useful too, and it very good
>>>> results.
>>>>
>>> yep
>>>
>>>  But not the best for 3d positioning of a point in space,
>>>>
>>> they will not provide any information regarding 3d positioning.
>>>
>>>
>>> the question is what information do you really want...
>>> if you want the position of someone in space, then you have to use a
>>> camera. but you can also track he's movement with a wacom.
>>> if you want to track it's gesture, then you need something faster. faster
>>> than a camera.
>>>
>>>
>>>
>>>  but excellent at giving the rotation of that point in space (i.e.:
>>>> accels, gyros).
>>>>
>>> yep.
>>>
>>> cyrille
>>>
>>>
>>>> :)
>>>>
>>>>
>>>> On Sat, Jan 8, 2011 at 7:21 PM, marcello <wallacechemical at gmail.com<mailto:
>>>> wallacechemical at gmail.com>> wrote:
>>>>
>>>>    On 08/01/11 14:25, cyrille henry wrote:
>>>>     > hello,
>>>>     >
>>>>     > for the kung fu : movement are usually very fast. a camera is
>>>> usually to
>>>>     > slow to be useful.
>>>>     >
>>>>     > i personally prefer using sensor like accelerometter on the body.
>>>>     > you can use wiimote or some other (faster) sensors system.
>>>>     >
>>>>     > here is a video of a performance i worked for few years ago.
>>>>     > http://vimeo.com/4596629
>>>>     > sensors are used at time : 4'57 to 6'06 and from 6'43 to 7'23.
>>>>     >
>>>>     > at 6'43, you can see that this kind of sensors are lot's more
>>>> accurate
>>>>     > that what you'll be able to do with a camera.
>>>>
>>>>    That's a good point about the sensors, I didn't know this
>>>> camera-latency
>>>>    problem.
>>>>    WHat kind of sensor have you used in your work ?
>>>>    Have you used wireless sensors ?
>>>>
>>>>     > imo, camera are good to track body position.
>>>>     > kinect look to have a sensing zone that is a bit limited.
>>>>
>>>>    A combination of the two could be an idea, but then it's complicated
>>>> to
>>>>    merge the data...
>>>>
>>>>    Thanks for the infos
>>>>
>>>>
>>>>    _______________________________________________
>>>>    Pd-list at iem.at <mailto:Pd-list at iem.at> mailing list
>>>>
>>>>    UNSUBSCRIBE and account-management ->
>>>> http://lists.puredata.info/listinfo/pd-list
>>>>
>>>>
>>>>
>>>>
>>>> --
>>>> Pedro Lopes (MSc)
>>>> contact: pedro.lopes at ist.utl.pt <mailto:pedro.lopes at ist.utl.pt>
>>>>
>>>> website: http://web.ist.utl.pt/Pedro.Lopes
>>>>
>>>>
>>>>
>>>> _______________________________________________
>>>> Pd-list at iem.at mailing list
>>>> UNSUBSCRIBE and account-management ->
>>>> http://lists.puredata.info/listinfo/pd-list
>>>>
>>>
>>
>>
>> --
>> Pedro Lopes (MSc)
>> contact: pedro.lopes at ist.utl.pt
>> website: http://web.ist.utl.pt/Pedro.Lopes
>>
>> _______________________________________________
>>
>> Pd-list at iem.at mailing list
>> UNSUBSCRIBE and account-management ->
>> http://lists.puredata.info/listinfo/pd-list
>>
>>
>
>
> --
> Jaime E Oliver LR
>
> www.jaimeoliver.pe
>
> 858 750 0924 (cel)
> 858 202 1522 (home)
>
> _______________________________________________
> Pd-list at iem.at mailing list
> UNSUBSCRIBE and account-management ->
> http://lists.puredata.info/listinfo/pd-list
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20110111/9f53a3df/attachment.htm>