Combined features of cubic B-spline wavelet moments and Zernike moments for invariant character recognition

In this paper a new method of combining cubic B-spline wavelet moments (WMs) and Zernike moments (ZMs) into a common feature vector is proposed for invariant pattern classification. By doing so, the ability of ZMs to capture global features and WMs to differentiate between subtle variations in description can be utilized at the same time. Analysis and simulations verify that the new method achieves better performance with respect to classification accuracy than using ZMs or WMs separately. In addition, this new method should also be applicable to other areas of pattern recognition.