Interest points

Local features, also called interest points, are characterized by sudden changes of intensity in the region. These local features are usually classified in edges, corners, and blobs. OpenCV encapsulates interesting point information in the KeyPoint class, which contains the following data:

• The coordinates of the interest point (the Point2f type)

• Diameter of the meaningful keypoint neighborhood

• Orientation of the keypoint

• Strength of the keypoint, which depends on the keypoint detector that is selected

• Pyramid layer (octave) from which the keypoint has been extracted; octaves are used in some descriptors such as SIFT, SURF, FREAK, or BRISK

• Object ID used to perform clustering