Penmanship Learning Support System: Feature Extraction for Online Handwritten Characters

Abstract : This paper proposes a feature extraction method for online handwritten characters for a penmanship learning support system. This system has a database of model characters. It evaluates the characters a learner writes by comparing them with the model characters. However, if we prepare feature information for every character, information must be input every time a model character is added. Therefore, we propose a method of automatically extracting features from handwritten characters. In this paper, we examine whether it correctly identifies the turns in strokes as features. The resulting extraction rate is 80% and in the remaining 20% of cases, it extracted an area near a turn.