“…For instance, Wenyin et al [12] introduced a spotting method including a learning step and an user feedback process. In [11,8,9], only symbols which satisfy some conditions (e.g.. convexity, connectivity or closure) are retrieved. A topology graph used in [4] for submatching or coarse specification of query provides good results; in [7], a modification of a shape representation based on Shape Contexts(called SCIP descriptor) is introduced in order to describe a graphic symbol using visual words to encode the image content.…”