Abstract

There are lots of digitized Chinese calligraphic works in CADAL calligraphic system. Since some calligraphic characters hadn't been written or though were ever written but were hard to read because of long time weathering. It is a challenging job to generate a piece of personalized calligraphic work with specific style. In this paper, we propose a novel Chinese calligraphic character synthesis method based on multigranularity deep search (MGDS). Moreover, an interactive polygon-based unit segmentation (IPUS) algorithm is proposed to separate radicals and strokes from calligraphic character images. Experimental results show that our proposed approach is effective and efficient for calligraphic character synthesis.

Paper Details