O'Reilly logo

Computer Vision with Python 3 by Saurabh Kapur

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

Detecting SURF keypoints

SURF keypoints are computed using concepts similar to Haar wavelets. Just like in SIFT, we did an approximation by using Difference of Gaussian instead of Laplacian of Gaussian. For SURF, we will use integral images (like for Haar Cascades) to speed up the keypoint detection step. SURF uses a technique called the fast Hessian detector that will be described next.

To select the location and scale of keypoints, SURF uses the determinant of the Hessian matrix. SURF proves that Gaussian is overrated as the property that no new structures can appear while going down to lower resolutions has only been proved in 1D, but does not apply to the 2D case. Given SIFT's success with the log approximation, SURF further approximates ...

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, interactive tutorials, and more.

Start Free Trial

No credit card required