My Question stands in the Title, before Answering please note, that I'm not referring to [this](http://answers.opencv.org/question/43674/sift-detector-returns-duplicate-keypoints/) and [this](http://answers.opencv.org/question/7701/why-sift-detector-returns-duplicate-keypoints/) kind of similarity.
I am asking that, because I see in the [Sift.cpp](https://github.com/opencv/opencv_contrib/blob/master/modules/xfeatures2d/src/sift.cpp) form contrib-master that it calls the KeyPointFilter function `removeDublicated(...)` in the [KeyPoint.cpp](https://github.com/opencv/opencv/blob/master/modules/features2d/src/keypoint.cpp) it shows that it only removes the one that are having the same Angle, Size or Coordinates.
So the only possibility, that the Filter would do anything is, when the KeyPoint is really a total Duplication. **But is this even possible in the SIFT Detection and Creation?**
I tested multiple example Images where it never created a Duplication (Testet through output of vector.size() after using `removedDuplicated(...)`
The Reason I ask is, because I want to remove every Code that is unnecessary from my own SIFT example, to make it as fast as possible.
↧