What is sift algorithm?

Sift algorithm is a local feature descriptor proposed by David Lowe in 1999, which was further developed and improved in 2004. Sift feature matching algorithm can deal with the matching problem when translation, rotation and affine transformation occur between two images, and has strong matching ability.

The inspiration of this algorithm is also intuitive. When human eyes observe whether two pictures match, they will notice typical areas (feature points). If this feature point region extraction process can be realized, and then the extracted region can be described, feature matching can be realized.

Application of sift algorithm

At present, SIFT algorithm has different degrees of application in military, industrial and civil fields, and its application has penetrated into many fields. Typical applications are as follows: object recognition; Robot positioning and navigation; Image mosaic; Three-dimensional modeling; Gesture recognition; Video tracking; Annotation identification; Fingerprint and face recognition; Feature extraction of crime scene.