A new methodology of online handwritten English character segmentation based on rules

This paper proposes an online handwritten English characters segmentation method based on rules. Firstly, all the local minima of the character coordinate sequences are treated as potential segmentation points; Secondly, we calculate 5 features of every potential segmentation point; Finally, the potential segmentation points are modified according to the rules to get the final segmentation points. The modification contains shift, deletion and maintenance. Thus, the coordinate sequences of characters are segmented to sub-sequences of characters by the final segmentation points. The experiment conducted on UNIPEN database shows that our method is feasible and effective.