United Zernike Invariants for Character Images

摘要

Feature extraction is one of the major components in traditional pattern recognition. There are many methods for extracting the features, either structural approach or global approach. In this paper, we present integrated formulation of Zernike Moments and United Moment Invariant for extracting the character images accordingly. The extraction values are validated by measuring the Inter-class and intra-class analysis to illustrate the effectiveness of the proposed solution. The results yield that the proposed method are feasible and better for extracting the images for both inter-class and intra-class analysis.