primitive object recognition (Hydro)

edit

Hi again,

in case that my other question regarding object recognition with only a 2d laser scan available cannot be answered easy enough...

How about any sort of object recognition using vision sensors like from Kinect through TOF cams and all the way down to simple web cams? Are there any close to ready built stacks/packages in ROS Hydromedusa that I could make use of? Up-to-date tutorials about that topic?

Thanks again!

Cheers, Hendrik