基于多阈值和多切分策略的间隙切分算法

摘要

Based on the analysis of traditional Chinese character segmentation, the thesis introduces clearance segmentation based on multi-threshold and multi-segmentation strategy. In the early stage of segmentation algorithm, implement segmentation through the characters spacing and multiple threshold to get the best results of the early segmentation. In the second segmentation stage, cut again on the best results of the early segmentation through multiple threshold. Experimental results show that, compared with the general clearance segmentation method, this algorithm can greatly improve the accuracy of segmentation of chinese characters.%在分析传统手写汉字切分算法的基础上,提出了基于多阈值和多切分策略的间隙切分算法.算法初切分阶段就字符间距通过多个阈值进行切分,从而得到预切分的最佳结果,在二次切分阶段,通过多个阈值对初切分阶段的切分字块进行再次切分.实验结果表明,该算法相对于一般的间隙切分法,对汉字切分的准确率有较大的提高.