Image Matching-SIFT

 

SIFT is short for Scale Invariant Feature Transform, which is a typical feature-based matching method. The basic principle of SIFT is to transform image data into scale-invariant coordinates relative to local features, find out the special points and use the extracted features to performance the matching.This method algorithm was published in 1999 and modified in 2004 by David G. Lowe.

 

SIFT Matching

 

 

 

WEB link: http://www.cs.ubc.ca/~lowe/keypoints/