while running haar classifier, how many times does the classifier run face detection on both videos and images?. for eg; with increasing scale range of 10% per loop.. i saw too many rectangles plotted on a single face when i gave 0 for drop grouping..