This paper proposes a model for recovering writing sequence from offline handwritten Chinese character image. In this model, a 4-layer hierarchy is presented to model Chinese character, where characters, components, subcomponents and strokes are located at each layer, respectively. Characters are decomposed to components and components are decomposed to subcomponents in turn by four decomposing operators. The totally-ordered relations among subcomponents are retrieved by defining the corresponding rules between decomposed relations and a partially-ordered set (poset) of subcomponents, and the writing orders of subcomponent are recovered by classifying strokes and crossing stroke pairs. Finally, experimental results show that our method is effective and accurate.
A Model for Recovering Writing Sequence from Offline Handwritten Chinese Character Image
2008 Congress on Image and Signal Processing ; 1 ; 298-302
2008-05-01
514624 byte
Conference paper
Electronic Resource
English
Discriminative Training for HMM-Based Offline Handwritten Character Recognition
British Library Conference Proceedings | 2003
|Handwritten Chinese Character Recognition: Alternatives to Nonlinear Normalization
British Library Conference Proceedings | 2003
|