Distinctive Image Features from Scale-Invariant Keypoints -- SIFT

The technique we are going to describe was proposed by D. Lowe in [25]. The following are the major stages of computation in image feature extraction:

• Scale-space peak selection: identify points of interest
• Keypoints localization: keypoints are selected based on measures of their stability
• Orientation assignment: one or more orientations are assigned to each keypoint based on local image properties
• Keypoint descriptor: local image gradients are measured at the selected scale in the region around each keypoint and transformed into a representation that allows for local shape distortion and change in illumination.

We made a simplification of this method. We treat every region of interest as a keypoint.