Chinese Calligraphic Character Retrieval Based on Shape Similarity

People identify individual characters mainly by their shape instead of texture or color.In this paper,a shape similarity based approach for Chinese calligraphic character retrieval is proposed.By the approach,first,the digitized pages of calligraphy works are segmented into individual characters,and then scaled to a standard size.Second,contour points are extracted to represent different calligraphic characters in shape.Third,similarity matching is made between the feature of the query image and the features of candidate calligraphic character images in the database. Finally,different styles of the same character are ranked and displayed according to the matching cost.Besides,the efficiency of our approach is illustrated by a preliminary experiment.