ABSTRACT
SIFT (scale invariant feature transform) algorithm was first proposed by Lowe [1] in 1999. This algorithm can extract the stable feature points, and deal with the matching problems of processing translation, rotation, ane transformation and viewpoint transformation between images. It is a kind of registration algorithms based on feature, and it has strong robustness which makes its own the best matching performance. What’s more, it is studied most widely currently.
