A computational evaluation system of Chinese calligraphy via extended possibility-probability distribution method

Robotic calligraphy has became a popular research topic in robotics. Therefore, a computational calligraphy evaluation system is required to access the quality of robotic writing results. This paper applies three types of feature criteria, derived from Chinese calligraphy theories, to extract features of Chinese characters from Chinese Calligraphy textbooks. Then, the Possibility-Probability Distribution method deals with these extracted features, so as to obtain the feature distribution of quality handwriting characters. The Possibility-Probability Distribution method uses the extracted features to automatically build an interior-outer-set computational model based on information diffusion theory. When the computational model is established, each Chinese character, written by a robot, is also extracted to three features; then, the computational model estimates each character's evaluation value. The experimental results demonstrate that the proposed method successfully produces an interior-outer-set computational model from Chinese calligraphy books. In particular, the model is able to generate an evaluation result for each character written by a robot system. To check the validation of the computational model, these characters are also evaluated by human experts. The comparison shows that the evaluation results of human experts are very similar to that of the computational model.