Velocity moments

Traditional moments (as applied to images), are designed to describe a single image. Velocity moments [3,4] are designed to compactly describe a shape and its motion through a sequence of images. There are currently two versions of the velocity moments: Cartesian and Zernike. The Cartesian velocity moments are based around the Cartesian moments which are non-orthogonal, but provide translation and scale invariance (provided the scale does not change within the sequence). The Zernike Velocity moments are based on the orthogonal Zernike moments producing a less correlated compact description, which is both translation and fully scale invariant (i.e. including scale changes within the sequence).