according to the docs the size is the diameter of the meaningful keypoint neighborhood . what does this exactly mean in the context of SIFT algorithm for example and when it is used ?
This forum is disabled, please visit https://forum.opencv.org