An Abstracting Method and Its Implementation for Feature-Points in Handwritten Chinese Characters Based on the Segmenting Technique

周昌乐; 张雄伟

您当前的位置：

首页 >

文章列表页 >

An Abstracting Method and Its Implementation for Feature-Points in Handwritten Chinese Characters Based on the Segmenting Technique

更新时间：2025-12-08

- An Abstracting Method and Its Implementation for Feature-Points in Handwritten Chinese Characters Based on the Segmenting Technique
- Acta Electronica Sinica Issue 5, (1997)
- 作者机构：
  
  1. 杭州大学计算机科学系
  2. 北京大学视觉与听觉国家实验室
  3. 杭州大学计算机科学系北京大学视觉与听觉国家实验室
- 作者简介：
- 基金信息：
- DOI：
  CLC： TP391.41
- Published：1997
- 稿件说明：
移动端阅览
[1]周昌乐,张雄伟.一种基于段化的手写汉字特征点提取方法及其实现[J].电子学报,1997(05):57-60.

周昌乐, 张雄伟. An Abstracting Method and Its Implementation for Feature-Points in Handwritten Chinese Characters Based on the Segmenting Technique[J]. Acta Electronica Sinica, 1997, (5).
[1]周昌乐,张雄伟.一种基于段化的手写汉字特征点提取方法及其实现[J].电子学报,1997(05):57-60. DOI：

周昌乐, 张雄伟. An Abstracting Method and Its Implementation for Feature-Points in Handwritten Chinese Characters Based on the Segmenting Technique[J]. Acta Electronica Sinica, 1997, (5). DOI：

摘要

手写汉字识别中的特征点提取，一直是结构匹配方法中的一个关键问题．系统识别率很大程度上取决于特征点提取的正确率．本文在总结前人和作者以前工作的基础上，提出了一种新的段化特征点提取方法，并具体用于手写汉字识别的实用系统中，取得了十分理想的效果．从已测试的１８００×１０个样本字的结果看，特征点提取正确率突破了９８％大关，为实用化手写汉字识别系统奠定了基础．

Abstract

It is a key problem to abstract the feature-points in the structure match methodfor recognizing handwritten Chinese characters

and the OCR system’s recognition correct rate depends on the correct rate of abstracting the feature-points to a great extent too. Based on the pastresearch works

we have provided a new method for abstracting feature-points and obtained quiteeffective results by applying this method to the practical system. The testing results show that thecorrect rate of abstracting feature-points in our method is up to 98% for 1800×10 sample characters

thus providing the fundament for building the practical system of handwritten Chinese characters.

关键词

Keywords

references

Views

241

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

No data

Related Author

No data

Related Institution

No data

⁰