Extracting the trajectory of writing brush in Chinese character calligraphy

This paper describes the extraction of the trajectory of the writing brush in Chinese character calligraphy (CCC), based on image and curve processing techniques and the calligraphy knowledge. This trajectory is used in a CCC robot which is developed to inherit CCC techniques. In CCC, the writing styles can be roughly classified into five different styles-ancient, angular, block, semi-cursive, and cursive style. This paper is limited to discuss the trajectory extraction from the character image written in block style. Firstly, for a given Chinese character, its image patterns in block style are retrieved from CCC database which contains 29,456 characters written by different famous calligraphers in Chinese history. Then the image of the designated writing is thinned. The coordinates of the line passing the centers of each stroke are detected from the thinned image with aid of writing order information. These coordinates represent the thinned-center-line of the stroke (TCLS, for short). And then, TCLS is separated into several curve segments according the calligraphy knowledge. The trajectory of the writing brush is considered as B-spline curves determined by the points on curve segments. The trajectory and the pressure control information are sent to the CCC robot to imitate calligrapher's behavior. The experiment results show that the proposed method obtains very good trajectories of the writing brush for CCC robot. (C) 2004 Elsevier Ltd. All rights reserved.