An evaluation on circularly orthogonal moments for image representation

Abstract

Invariant image representation plays an important role in many pattern recognition applications, such as texture classification, face recognition and character recognition. In this paper, we evaluate some of the invariant orthogonal moments, including Zernike moment (ZM), pseudo-Zernike moment (PZM), and Polar Harmonic Transform (PHT), which are all computed in a circular domain. The performance of invariance and image representation ability are evaluated.