1 | initial version |
You should check out the cascade classifier documentation. It has an examplr of how to load and use a cascade to detect an actual object. However, if you are looking for a fast pedestrian detector, this one won't satisfy your needs either. You should look into class specific hoigh forests or grammar based latent SVM detection. Also the integral channel processing is a promising technique.
However many of these techniques require you to probram your own code based on provided sub functionality of openCV.