A Quick Search Engine for Historical Chinese Calligraphy Character Image

Libraries and museums are digitizing their collections of historical culture objects to enable public access, such as historical Chinese calligraphy. These collections are only available in image format, lacking practical technology to offer the basic search service for public access. This paper proposes a quick search approach by a coarse-to-fine strategy. First, long list of calligraphy characters are pruned into shorter list by eliminating characters that have no possibility to be similar to the query. Then shape matching algorithm is employed to quantitatively measure the similarity between the query and each remainder calligraphy character. Finally, the efficiency of the algorithm is demonstrated by experiments with 12,021 images of calligraphy character.