摘要
一直以来,让机器具有模式识别能力一直是计算机科学家们的努力方向。研究模式识别,是理解人类智能的本质的重要途径。字符识别是一个典型的模式识别问题,手写数字识别具有不同字符体型相差不大,相同字符有多种不同写法,数字没有上下文关联等特点,这些特点使手写数字识别比较困难。该论文设计了一套基于KNN算法,用Python语言实现手写数字识别系统。先对图片进行预处理,预处理包括灰度化、去噪、二值化等等,再用KNN算法对图片进行分类,手写数字识别也是一个十分类的问题。实验结果表明该文所设计的手写数字识别具有较好的识别率,同时也说明KNN算法在手写数字识别上具有良好的应用。
All along, There always been a direction of computer scientists＇ efforts of Making the machine a pattern recognition capability. Research pattern recognition is an important way to understand the essence of human intelligence. Character recogni- tion is a typical pattern recognition problem. Handwritten digital recognition has different characters. The same characters have many different Wordings, and the numbers are not context-sensitive. These features make handwritten digital recognition diffi- cult. This paper designs a set of KNN algorithm Based on Python language to achieve handwritten digital recognition system. First of all, the image preprocessing, including gray, denoising, binarization, etc., and then use KNN algorithm to classification image, the handwritten digital recognition is a ten classification problem. The experimental results show that the handwritten digi- tal recognition in this paper has a good recognition rate, and also shows that KNN algorithm has a good application in handwrit- ten numeral recognition.
作者
李诗语
王峰
曹彬
梅琪
肖飞
LI Shi-yu, Everett X. WANG, CAO Bin, MEI Qi,XIAO Fei （ School of Infornlation Engineering, Guangdong University of Technology, Guangzhou 510006, China ）
出处
《电脑知识与技术》
2017年第9期175-177,共3页
Computer Knowledge and Technology
基金
基于累积展开方法的纳米电子器件非平衡输运和量子效应理论研究（2016A030313698）
移动互联网协助的高精度北斗卫星导航增强系统（20148090901070）
基于光载无线技术的舰艇物联网平台的研究开发（20158090901060）
广州市科技计划项目（201604046007）
广东省省级科技计划（产学研）项目（20168090918031）
关键词
KNN算法
PYTHON语言
模式识别
预处理
手写数字识别
KNN algorithm
Python language
pattern recognition
preprocessing
handwritten numeral recognition
