On Moment Invariants Based Seal Verification Technology

Seal verification plays an important role as an entity identification method. The traditional manual method of seal verification is restricted by subjective experience and external factors with lower efficiency and unavoidable mistakes. Thus, the study of automatic seal verification technology has wide application value and important theory meaning.In this paper, a complete seal extracted scheme, with the combination of the RGB colure model and the HS V colure model, is proposed. The seal is extracted by analyzing the RGB model and the background is extracted by using the HSV model. Then the missed part can be reconstructed by associated the seal with the background. The experiment result indicates a better effect.A new verification method is proposed in this paper on the basis of Zernike moment invariants which is with the relation of moment theory. The Zernike moment invariants of the seals are extracted as features after the normalization of the seals translation and scaling. Then the seals are classified and verified by using the weighted Europe distance. The proposed method can be implemented without rotation compensation and is unrestrained for shape of seals. Compared with the Hu moment invariants based seal verification method, the recognition rate of this method is improved greatly.