Determining the Identity of Author by Combining Wavelet Transform and Gabor Filter

It’s a long time that researches in the field of identifying the author, based on manuscripts, is carried out and in most parts of the world is almost done. Using intelligent computer systems in this area is relatively new and has been done since 2222. In this paper, an offline text-independent method for determining the identity of the author, using handwritten text is proposed. Based on previous studies, handwritten text is taken into account as a unified image. After several steps including preprocessing, using Gabor filters and wavelet transform features are extracted. These features, after analyzing, by the k-nearest neighbor classifier, are applied to identify the author based on the available manuscripts that have already been used. In addition, a new method for extracting features from the output of Gabor filters is proposed which is based on characteristics of the horizontal and vertical gradient energy and the human visual system. Results of implemented method son a standard database image admit that the proposed method has a good efficiency compared with existing methods which are often very complex and time consuming.