Chinese character recognition by Zernike moments

Moment descriptors have been applied in object recognition as the features since the moment method was introduced by Hu [1]. The moment based features capture the global properties of an object rather than the local ones. In this research, a set of Zernike moment based feature vectors is proposed for a Chinese characters recognition system. We have composed three different feature vectors in the four-dimensional Zernike moment space by evaluating the variance values of lower order Zernike moments with encouraging experimental results. We have also clarified the invariant properties of Zernike moments in our system.