The last days i was looking for another application for doing the computervision job, since eyesWeb’s “extract regions” block swaps the recognized blobs’ IDs when they change their position, which is problematic when one wants to track, or follow more than one moving object. Christian Iten and Daniel Luethi from the HGKZ told me that there was an Object for Max/MSP/Jitter called 2up.jit.centriods that can already track the spots. But Christian (Riekoff, not Iten) and me agreed to keep the multitouch screen as OpenSource as possible, so i ended up using the powerful freeware tool vvvv.
i have become a true fan in no time, since there are many advantages:
- its’s free
- it’s fast
- its interface, once understood, is the most minimal but effective i’ve ever seen.
- the “Contour” node (thats what Objects or Blocks are called in vvvv) gives us width, height, coordinates, form (sub-coordinates of border points) and orientation plus a Permanent ID of each blob.
right now i am rewriting (since everything is done by vvvv, this means throwing out wannabe-code) the OSCtracker in Processing. may i still call it “tracker”?