An integrated formulation of Zernike representation in character images

Abstract

Many studies have been done on improving Geometric Moment Invariants proposed by Hu since 1962 for pattern recognition. However, many researchers have found that there are some drawbacks with this geometric moment. Hence, this paper presents an integrated formulation of United Moment and Aspect Moment into Zernike Moment Invariant to seek the invarianceness of the solutions. The proposed method will be validated mathematically and experimentally. The validity invarianceness of the proposed method is measured by conducting the intra-class and inter-class analysis. The results of the proposed method are promising and feasible in identifying the similarity and differences of the images accordingly. © 2010 Springer-Verlag.