Hello,
I am using the BOW approach to solve object detection for a large library of image types.
http://www.morethantechnical.com/2011/08/25/a-simple-object-classifier-with-bag-of-words-using-opencv-2-3-w-code/
The problem i have is that sometimes more than one object can appear in the scene. So if the image contains two objects of different types what is the approach to separate the features and detect the objects individually.
Regards,
Daniel