Online Diagramming Recognition Based on Automatic Stroke Parsing and Bayesian Classifier
Qiang Xie, Zhengxing Sun, Guihuan Feng · 2006
Owing to the fluent and lightweight nature of freehand drawing, sketch-based user interface is becoming increasingly significant in graphical computing. However, it is still an important problem that user can fluently draw with such a tool. This paper presents a strategy for online diagramming recognition. Two distinct characteristics are addressed. Firstly, a method of automatic stroke parsing is proposed based on spatial proximity parameter. It can automatically and quickly group inputting strokes into a single object as user intended. Secondly, a Bayesian classifier is designed to implement domain-independent online diagramming recognition. It can recognize the inputting symbols insensitive to drawing styles of different users with high precision. The experiment results prove the effectiveness and fluentness of our prototype system for different users