OpenCV FAST operation. Why use FAST instead of FastFeatureDetector

I have been experimenting with the FAST algorithm (in both CPU only and GPU versions) using the class cv::FastFeatureDetector (or the cv:cuda:FastFeatureDetector )

Then I found my colleague is using cv::FAST

The first one is a “Wrapping class for feature detection using the FAST method” while the second one is a function that Detects corners using the FAST algorithm

My questions are two:

  1. What could be the criteria to use one over the other?
  2. Is there a CUDA alternative to use instead of cv::FAST ?

the difference is only in the interface
( FastFeatureDetector::detect() calls FAST() internally)

1 Like