摘要： 光学字符识别技术在人机交互中，发挥了界面友好的作用。根据识别时所提取具体特征的不同，字符识别通常可以分为：统计字符识别、结构字符识别和神经网络字符识别技术3种。该文分析了文字识别中的典型方法，并结合模式识别领域取得的理论成果，介绍了模糊技术识别、基于语义理解的识别等发展中的识别技术。光学字符识别为大量数据的自动录入提供了一条有效的途径，在文档自动分析、图像检索、人机交互方面有广阔的发展空间。

关键词: 光学字符识别, 特征提取, 模式匹配, 结构分析

Abstract: Optical Character Recognition acts as a friendly interface between human and computer.According to the different methods of feature extraction，there are 3 categories—statistic analysis,structure analysis and ANN classification method.This paper analyzes some of the classical methods.Some new solutions such as fuzzy analysis and context analysis are also introduced.OCR contributes a lot to the automatical conversion of large volumes of data.In the future,however,there are a lot to be done in the field of document analysis,image query and human-computer interaction.

中图分类号: 

• TP391.41