Writer identification using global wavelet-based features

In the human society, it is very important to find out the true writer of an unknown handwriting document. Therefore handwriting-based writer identification has been a hot research topic in pattern recognition field since several decades before. In our research, we find that the global styles of different people's handwritings are obviously distinctive and the histogram of the wavelet coefficients of preprocessed handwriting image can be well characterized by the generalized Gaussian model (GGD) in wavelet domain. As a consequence, in this paper, we propose a new method by combining wavelet transform and GGD model for writer identification of Chinese handwriting document. Tested by our experiment, this method achieves a satisfied identification result and computational efficiency as well.