Multiple object class detection by extracting image features (like SIFT or HOG)

edit

Hello,

I am trying to do object detection on objects on the road (cars, humans). However I do not want to use deep learning approach and want to extract image feature points and pass it to a classifier. However, after reading and searching a lot, I find mostly all such methods are for a single class. Please let me know if anybody has any lead on this?

Your help would be much appreciated.

thanks