1) Which version of Processing are you going to be using?

2) Are you willing to have some calibration steps involved, in order to reduce false positives ?

For Kinect, I would look at SimpleOpenNI. 

For openCV, that depends on whether you are working on 1.5.1 or 2.0aXX ... 

But I would recommend Blob Scanner, as it runs on both. 

Haven't used it too much yet though... I've used blobDetection quite often, and it's quite easy to use... 

Hope that helps... 

~ J

1. I have P1.5.1 installed and already loaded with minim for a simple pilot project which is working super fine. That installation seems to have NOT uninstalled 1.0.1 which is the one I had before it, but ofc. would try to keep this as current as possible (so 1.5.1 preferred).

2. Sure, a 1 min calibration before performance shouldn't be a show stopper.  Dream use case is being able to project and perform with this code during a live show of my music.

I could get into buying the Kinect hardware at a later stage, when I know for sure the investment is vetted by the already mostly-worked out code, but atm I would try to develop the code using my built-in web-cam, or even some low-res cheapo that somehow landed on my gadget drawer.  Does SimpleOpenNI work with a plain vanilla web cam?

What I take away from your post is that what I want to do falls within what CV people call blob-detection; If that is correct then at least that narrows the possible libraries down a lot.  So many thanks is what I'm saying.

Please expand or correct me as needed, tnx.

-h.

1) That's fine, 1.5.1 will work with all of this, so no worries there. 

2) In your case, it will likely be a simple adjusting of thresholds that you could adjust via keys or a gui slider, etc.

3) SimpleOpenNI only works with Kinect (not quite true, but it doesnt work with basic webcams...)

4) Blob Detection for sure... Id start there, youll get some functionality and learn the ropes. You may want to move into HAAR Cascade, which will let you identify faces, features, gestures, etc. later...

Many thanks for sorting out this n00b over here.

I wish you well,

-h.