The recognition method for roughly hand-drawn logical diagrams based on hybrid utilization of multi-layered knowledge

A recognition method for roughly hand-drawn diagrams based on hybrid utilization of multilayered knowledge is described. This method is free from various drawing constraints which previous systems cannot avoid. It is based on two concepts: (1) no strict segmentation should be performed-various hypotheses of symbols should be considered; and (2) various kinds of knowledge should be utilized-ambiguity of a figure can thus be resolved. Experimental results show that this method is effective for the recognition of roughly hand-drawn logical diagrams.<<ETX>>