findbounds and cv.jit for face-tracking


    Aug 26 2013 | 11:17 am
    I hope to set up a system so that when a person sits in front of a screen, their face is 'automatically' detected, and the x/y coordinates of the face provide spatial data for some Open GL objects in 3D space. There are two problems -
    - if I use jit.findbounds to locate 'face toned' information, the sensitivity seems to be too acute to be reliable for different faces and light conditions.
    - the CV objects appear to need user input - a mouse click - to designate the locus of tracking. I'm hoping for a system that will pickup faces (or body at least) automatically.

    • Aug 27 2013 | 8:28 pm
      @Brucem:
      have you checked [cv.jit.faces]? If I remember correctly, it doesn't need user input to work