You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi,
I'm using latest stable OpenCV (not v3) and trying out FREAK. I would like to use the special matching that is used in the demo. I was previously using the BF matcher in the following way:
However, it does not seem to be working. I don't think the HammingSeg operator() is even being called. I tried printing something to the screen from there and it does not seem to be reaching the code.
Do you know if this should work like this?
By the way, when using this custom matcher, the idea is not to require the standard ratio-based test between 1st and 2nd neighbors, right?
EDIT: I checked the demo, and placed the print line on the operator() function and it does not seem to be using this either. I'm suspecting it uses standard HAMMING based BFMatcher.
The text was updated successfully, but these errors were encountered:
Hi,
I'm using latest stable OpenCV (not v3) and trying out FREAK. I would like to use the special matching that is used in the demo. I was previously using the BF matcher in the following way:
matcher = boost::make_sharedcv::BFMatcher(cv::NORM_HAMMING);
Later, I add descriptors using add(). Finally, I match using knnMatch().
To try your custom matcher I replaced the first line with:
matcher = boost::make_sharedcv::BruteForceMatcher<cv::HammingSeg<30,4>>();
However, it does not seem to be working. I don't think the HammingSeg operator() is even being called. I tried printing something to the screen from there and it does not seem to be reaching the code.
Do you know if this should work like this?
By the way, when using this custom matcher, the idea is not to require the standard ratio-based test between 1st and 2nd neighbors, right?
EDIT: I checked the demo, and placed the print line on the operator() function and it does not seem to be using this either. I'm suspecting it uses standard HAMMING based BFMatcher.
The text was updated successfully, but these errors were encountered: