A Novel Method for Improving Artifacts of Chinese Calligraphy Character Skeleton Extraction

Chinese calligraphy style evaluation and generation are important to the aim of cultural heritage preservation. Character skeleton extraction is a key preprocessing step and most existing algorithms suffer from drawbacks of non-single pixel width, bifurcation etc. In this paper, an optimization process is proposed to overcome these drawbacks to some extent. It gets the provisional skeleton using a rotation invariant thinning algorithm, and proposes a two-step process to further optimize the results: a) to reduce the skeleton to single pixel in width, and b) a maximum circle method to merge the bifurcation points. Experimental results on several databases show that the proposed approach obtains better skeletons than existing benchmarking methods.