Ask Your Question

Which feature descriptor should I use with Harris corner detector?

asked 2012-10-19 17:45:20 -0500

valgussev gravatar image

updated 2012-10-22 15:05:39 -0500

I'm new to OpenCV, having a question: I'm interesting in this tutorial - Features2D + Homography to find a known object, as I can see it uses SurfFeatureDetector and then SurfDescriptorExtractor. If I will use the Harris corner detector which DescriptorExtractor I should use?

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2012-10-21 10:01:28 -0500

alexcv gravatar image


A classical combination is Harris detector+SIFT descriptor, you can also consider OpponentSIFT is color information is an important/selective criteria.


edit flag offensive delete link more


Thank you! So if i'll just find corners with Harris detector and then will use SiftDescriptorExtractor it's gonna work? Are "corners" and "features" in this case the same thing?

valgussev gravatar imagevalgussev ( 2012-10-21 11:23:54 -0500 )edit

Hi, actually not, Regarding the detector aspect, Harris and SIFT differ from the threshold value computation (hessian matrix vs second moment criteria). Regarding the SIFT description : descriptor is generally extracted from the same scale as its SIFT keypoint scale. It takes into account magnitude and orientation of local information computed on a grid of small patches around the keypoint. You should take a look at Lowe's paper (1999) <> and also compare to other descriptors GLOH, etc.) .


alexcv gravatar imagealexcv ( 2012-10-21 12:04:53 -0500 )edit

As I can see in this tutorial SurfDescriptorExtractor has just information about coordinates of the keypoints and then by itself calculate the descriptors. After the Harris corner detector I will have the same coordinates of the corners, so basically it should be the same. Am I wrong?

valgussev gravatar imagevalgussev ( 2012-10-21 14:25:59 -0500 )edit

Question Tools


Asked: 2012-10-19 17:45:20 -0500

Seen: 2,718 times

Last updated: Oct 22 '12