Shape recognition

did somebody realize a shape recognition patch (as an example of gestures out of a video or jpgs)? is it possible to use detect object for that?

max/msp/jitter seems to use also the lib for that purposes, does anybody know if this is the same stuff like detect object?

thanks for advice …

i’d guess that shape-recognition is rather done with something similar to the contours node. and i remember from browsing through opencv-documentation that there are some dedicated functions for comparing different contours (i.e. shapes).

so if your are a little into programming it shouldn’t be hard to hakc something together. starting at HowTo FreeFramePlugins.