CASIA OpenIR  > 毕业生  > 硕士学位论文
联机手写全屏汉字分割与识别
肖志宏
Subtype工学硕士
Thesis Advisor刘迎建 ; 李明镜
1998-12-01
Degree Grantor中国科学院自动化研究所
Place of Conferral中国科学院自动化研究所
Degree Discipline模式识别与智能系统
Abstract在计算机中文信息处理中,汉字输入是个重要问题,已有很多键盘输入编 码方式。联机手写输入是重要的非键盘输入方式,其中涉及到两个问题:方便 的用户界面,及准确的汉字识别方法。针对这两个问题,本文讨论了用户界面 中全屏字符分割问题,及汉字识别方法。 联机手写输入界面中,在全屏范围内书写的方式接近于一种自由书写状态, 其问题在于确定每个字的开始与结束,即把各字的笔画串相互区分开。本文结 合我们的产品研究项目,针对全屏手写字符分割问题,给出基于汉字书写结构 特点和笔画位置关系的分割方法:首先用简单的规则把笔画串分为基本的小块, 每个小块是单字或字的一部分,再由合并及评价准则通过反复的合并与分裂过 程,产生适当的分割结果。在实验中取得较好的效果。 联机手写汉字识别方法是联机手写输入的核心问题。多重相似度(MS)方法 最初应用于印刷体字符识别及限制性手写字符识别。该方法针对模式匹配方法 受形变影响很大的缺点,把一字的样本特征矢量集合在空间中的分布结构用几 个主轴方向来描述,用多重相似度的概念表达输入字符与一个字的相似性。本 文讨论了把它应用于联机手写汉字识别的问题。由于联机手写汉字形变太大, 实际效果不太好。
Other AbstractChinese character input method is becoming increasingly urgent in Chinese character information processing around the world. To do with the problem, many researchers have made great advances during the last decades of years. On-line handwritten character input is an important method in the area. In general, on-line handwritten approach involves two problems: convenient interactive operating interface, and the character recognition algorithm. This paper discusses a character segmentation method of the screen-wide writing interface, and the on-line character recognition method. Screen-wide handwriting fashion is close to a free handwriting method. The first problem is how to pick out the strokes of each character, which is demanded for the recognition algorithm. In the paper, we give a character segmentation method based on character structure and stroke position to segment the strokes into proper groups through combination and division circulation, which meets the demands of the practice. An other method based on recognition which needs further research is discussed in this paper, too. On-line hand-written character recognition algorithm is the main problem in the system. Multiple similarity method (MS) is a recognition method which was used in printed character and hand-printed character recognition. As the conventional pattern matching method is weak to position deviation and deformation of the input characters, MS method presents a new way to define the similarity between the input character and the patterns of a character available This paper discussed the application of the MS method as an on-line handwritten character recognition method. The result is not quite exciting yet. The feature extraction method is of great importance to the algorithm.
shelfnumXWLW498
Other Identifier498
Language中文
Document Type学位论文
Identifierhttp://ir.ia.ac.cn/handle/173211/7242
Collection毕业生_硕士学位论文
Recommended Citation
GB/T 7714
肖志宏. 联机手写全屏汉字分割与识别[D]. 中国科学院自动化研究所. 中国科学院自动化研究所,1998.
Files in This Item:
There are no files associated with this item.
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[肖志宏]'s Articles
Baidu academic
Similar articles in Baidu academic
[肖志宏]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[肖志宏]'s Articles
Terms of Use
No data!
Social Bookmark/Share
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.