Training new LatentSVMDetector Models.
Hi,
I haven't found any method to train new latent svm detector models using openCV. I'm currently using the existing models given in the xml files, but I would like to train my own.
Is there any method for doing so?
Thank you,
Gil.