In OpenCV 3.0 face detection demo, there are 4 options of detecting function:
CASCADE_FIND_BIGGEST_OBJECT
CASCADE_DO_ROUGH_SEARCH
CASCADE_DO_CANNY_PRUNING
CASCADE_SCALE_IMAGE
But I found that, only CASCADE_SCALE_IMAGE could be enabled, other methods have been removed from 3.0 beta version.
But, in previoius version of OpenCV, say 2.4.1, every method could be enabled. I found that scaling image is faster than scaling cascade, but the paper these methods based on--Rapid Object Detection using a Boosted Cascade of Simple Features, said scaling cascade is much faster. I wonder why this paradox happened? And why OpenCV new version remove the function of scaling cascade? If someone knows, please kindly tell me