SURF, SIFT? What about ORB?

When they first appeared, algorithms which detected feature detectors/descriptors were a great idea. In 1999 came SIFT (Scale-Invariant Feature Transform). In 2006 came SURF (Speeded Up Robust Features) which was suppose to be several times faster than SIFT. These algorithms are used for a huge range of computer vision and image processing applications : panorama stitching, object recognition, gesture recognition. The problem with both these algorithms is that the are patented, which makes them harder for the broader community to use. It also makes it harder to build into packages such as OpenCV.

A good alternative is ORB, short for Oriented FAST and rotated BRIEF [1], first presented in 2011 to provide a fast and efficient alternative to SIFT or SURF. It’s also available in OpenCV, and doesn’t have a patent. Works really nicely. It’s basically a blend of FAST keypoint detector and BRIEF descriptor with performance enhancements. I used it to match the Prokudin-Gorskii R-G-B images. Works very well. I’ll write a post on it next week.

Other feature detectors in OpenCV (3.X) include:

  • BRIEF ( Binary robust independent elementary features) [2] – The first binary descriptor published (only a descriptor).
  • BRISK (Binary robust invariant scalable keypoints) [3]
  • FREAK (Fast retina keypoint) [4] – Can be found in OpenCV xfeatures2d, as an experimental algorithm.
  • KAZE (a Japanese word the means wind) [5] –  (Also has accelerated KAZE, AKAZE).

[1] Rublee, E., Rabaud, V., Konolige, K., Bradski, G., “ORB: an efficient alternative to SIFT or SURF“, in IEEE Int. Conf.  on Computer Vision, pp. 2564-2571 (2011).
[2] Calonder, M., Lepetit V., Strecha C., Fua P. , “Brief: Binary robust independent elementary features”, in  Computer Vision, Springer, pp.778-792 (2010)
[3]  Leutenegger, S., , Chli, M., Siegwart, R.Y., “BRISK: Binary robust invariant scalable keypoints.”, in IEEE Int. Conf.  on Computer Vision (2011)
[4] Alahi, A.,  Ortiz, R., Vandergheynst, P.,  “Freak: Fast retina keypoint”,  in IEEE Int. Conf.  on Computer Vision and Pattern Recognition (2012)
[5] Alcantarilla, P.F., Bartoli, A., Davison, A.J., “KAZE features”, in European Conference on Computer Vision (2012)


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.