detectMultiScale in python vs C++

asked 2013-11-29 11:14:42 -0500

iLucian gravatar image

Running a cascade classifier, and the same trained cascade xml file I get different results in python than in C++.

I tried leaving all the C++ arguments as default, and I checked the detectMultiScale the file build/modules/python/pyopencv_generated_types.h, but still I can't get it to return the same results. I'm not familiar with Python OpenCV wrapper, is there anywhere I can read up on it? Or if someone has some advice on the subject, I'd be appreciative.

edit retag flag offensive close merge delete


adding some test case would be quite helpful.

berak gravatar imageberak ( 2013-11-29 12:58:49 -0500 )edit

I have the exact same problem. The python detectMultiScale gives slightly different results from C++... I wrote a prototype in Python, but when translating to C++ the detecton is not as good.

Joakim gravatar imageJoakim ( 2014-06-05 18:50:36 -0500 )edit