论文讨论了一种通过自然语言语义层次的理解来协助提高脱机手写体识别率的方法。从识别的过程来看,这种方法是一种基于上下文关系的识别“后处理”方法。支中介绍了语义单元、语义单元表示、语义单元表示树、语义单元表示集、语义单元表示树集,并且具体讨论了通过语义单元表示树集的剪枝来提高识别正确率的方法。
In this paper,we discuss through semantic analysis.Seeing from a new method to effectively improve offline handwritten character recognition the process of recognltion,this method is at post-processing stage based on context.We introduce SE (Semantic Element), SER (Semantic Element Representation), SER-tree, SER-Set, SER-treeSet and discuss this new method of pruning on the SER-treeSet in details.