The handwriting identification of copyright owners based on fractal features

T handwriting identification plays a very important role in the protection of copyright. This paper proposes a very efficient method to realize the automatic identification of handwriting. According to it the handwriting image of Chinese is projected separately on the X-axis, Y-axis, plusmn45degline, which are corresponding to horizontal line, top-down vertical line, left-downward slope line, short pausing stroke of Chinese characters. Then the projected curves on the four directions are obtained and the fractal dimension of these curves are extracted as the digital features of the handwriting image. With these features and the nearest neighbourhood algorithms, the writer identification is confirmed easily. The result of experiments shows that the methods proposed in this paper are valid.