Automatic Creation of Artistic Chinese Calligraphy

摘要

We present an approach to creating Chinese calligraphy with particular style from learning author's written works automatically. Our first contribution is a method to represent Chinese character topology via fuzzy relative position. Our second contribution is an algorithm to take topological features as features fed into the evaluation model, which is a decision tree algorithm. Our third contribution is an algorithm of heuristic search in the optimization glyph phase. To improve the converge speed, we introduce hypothesis testing and the decay function of transformation amplitude. The experiments demonstrate our algorithm can obtain the similar style Chinese calligraphy with training samples.