Hi
I've been looking for some documentation of how are the FREAK descriptor pattern is build, but neither the original paper nor the OpenCV implementation have information more than the implementation itself.
My question is if someone know or have information about how is this defined?