识别和几何信息融合的维吾尔文联机手写单词分割

摘要

通过对手写维吾尔文字中的字母连接特点的深入研究,提出了一种有效的基于动态规划的联机手写单词分割方案:首先,去掉单词中的附件部分后,通过分析主要笔划书写轨迹的形状,找出潜在的过分割点并合并被切分成的基本块与对应它的附加部分,得到基本字母片段序列；然后,对相邻的基本片段进行组合形成切分候选网格；再利用单字母分类信息和基于切分块的几何信息进行融合,采用动态规划算法来进行评价分析,从而动态分割并寻找出最优的分割路径.对于联机单词样本进行的实验证明,上述方案对于维吾尔文单词的分割有很好的效果.%Considering that correct and efficient segmentation of Uyghur words into characters is crucial to the successful recognition of Uyghur words,this paper presents a novel character segmentation method using dynamic programming in online cursive Uyghur handwriting to enable many connected characters separation in cursive Uyghur handwriting.The method consists of the steps below.Firstly,after removing delayed strokes from the handwritten words,potential breakpoints are detected from concavities and ligatures by temporal and shape analysis of the stroke trajectory,and reconstruct delayed strokes and a sequence of primitive segments are obtained.Then,by merging the neighboring blocks,candidate segmentation paths are created.Further,paths are evaluated by the character recognition and geometric information,and a dynamic programming algorithm is used to find the best segmentation point for each character.The preliminary experiments on an online Uyghur word dataset demonstrate that the proposed method can achieve good performance in segmenting cursive handwritten Uyghur characters.