I'm trying to match a picture saying "this is ducks, this is towers" by giving votes from the hough transform of a picture.

I've found this article http://ttic.uchicago.edu/~smaji/papers/hough-cvpr09.pdf, but I didn't find a implemented code (and I'm trying several things).

Do you know anything similar to that, for example included on OpenCV?



1 Answer 1


Hough forests for object recognition are initially proposed by Jurgen Gal. The paper was named "Class Specific Hough Forests for Object Recognition".

Here is his homepage. Here are the source code and the details. I think you could make use of those.

Certainly better methods exist and one could well talk about them. However I am sticking to the Hough framework, as you mentioned that in the first place.

  • $\begingroup$ Thank you. I'm not interesting in detecting objets, but seeing how close it is. On the webpage there's a code called "Scalable Multiclass Object Detection" but the code is broken. That's what I'm looking for, calculating the image number and compare it with the previous numbers that I've calculated before... do you know anything about that code? Thanks. $\endgroup$ Dec 15, 2014 at 12:26
  • $\begingroup$ I don't think it is broken. It is better if you contact the author for further details. $\endgroup$ Dec 15, 2014 at 14:31
  • $\begingroup$ files.is.tue.mpg.de/jgall/projects/houghMC/houghMC.html For contacting, it's broken as well :\ $\endgroup$ Dec 15, 2014 at 14:33
  • $\begingroup$ You can contact Juergen here: files.is.tue.mpg.de/jgall $\endgroup$ Dec 15, 2014 at 15:52

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.