no code implementations • 24 Jun 2014 • Moria Tau, Tal Hassner
Doing so allows scale invariant descriptors to be extracted anywhere in the image, not just in detected interest points.
