Calligraphy Character Synthesis from Small Sample Set
A novel approach to synthesize calligraphy characters is presented in this paper. Only a small set of calligraphy characters written by the specific calligrapher is needed. A robust polygon based radical and stroke extraction method is introduced, which can generate strokes and radicals preciselywith a few manually marked pixels. A new radical and stroke selection method called component selection algorithm is described, which can decide whether to use radicals or strokes and can find out the most suitable ones from the candidate radicals and strokes.After putting the radicals and strokes together and form the calligraphy character, the style difference among them must be minimized. In order to do this better, a new way to adjust the stroke widths is presented. A sample set containing only 30 selected calligraphy characters are used to synthesize new calligraphy characters. The results show that our approach works effectively.
Keywordscalligraphy character synthesis component selection algorithm small sample set
Unable to display preview. Download preview PDF.
- 4.Yu, K., Wu, J., Zhuang, Y.: Style-Consistency Calligraphy Synthesis Systemin Digital Library. In: 9th ACM/IEEE-CS Joint Conference on Digital Libraries, pp. 145–152. ACM Press, New York (2009)Google Scholar
- 6.Xu, S., Jiang, H., Lau, F.C., Pan, Y.: An Intelligent system for Chinese Calligraphy. In: The 22nd Conference on Association for the Advancement of Artificial Intelligence, pp. 1578–1583. AAAI, New York (2007)Google Scholar