摘要
字符分割的准确率直接影响到小学生手写口算题自动批改的结果。针对小学生手写口算题字符存在粘连的情况,本文提出了多阶段字符分割的方法。首先采用投影法对字符块进行分割;然后根据字符块宽度与笔画复杂程度,找出存在粘连的字符块;最后对滴水算法的水滴起始点进行改进,针对传统滴水算法起始点不准确和字符分割断裂的问题,细化粘连字符块并确定字符粘连点。实验结果表明,该方法可以有效分割手写口算题字符。
The accuracy of character segmentation directly affects the result of pupils'handwritten verbal arithmetic problems correction.To solve the problem of connected characters in handwritten verbal arithmetic problems,a multi-stage character segmentation method is proposed.First,the character blocks are segmented by projection method.Second,according to the width of the character block and the complexity of the stroke,the connected character blocks are found.Finally,the starting point of drop-falling algorithm is improved.To solve the problem of inaccurate starting point and broken character segmentation of the traditional drop-falling algorithm,the connected character block is thinned and the character connected point is determined.The experimental results show that this method can effectively segment characters of handwritten verbal arithmetic problems.
作者
贾华颖
颜佩珊
仲昭婷
梁银
JIA Huaying;YAN Peishan;ZHONG Zhaoting;LIANG Yin(Department of Computer Science and Technology,Jiangsu Normal University,Xuzhou,China,221116)
出处
《福建电脑》
2022年第12期30-33,共4页
Journal of Fujian Computer
基金
江苏省高等学校大学生创新创业训练计划项目(No.202010320131Y)资助。
关键词
粘连字符
滴水算法
投影法
字符分割
Connected Characters
Drop-falling Algorithm
Projection Method
Character Segmentation
