(毕业论文35页17129字)摘 要:手写体汉字识别是汉字识别领域中最难的课题。其应用于自然人机交互领域,提高了人机交互的自然性和友好性;同时也用于文字信息自动处理领域,为节省人力,提高效率,加快信息流动作出了巨大贡献。本文对目前手写体汉字识别在预处理、特征提取、分类识别及后处理四个阶段主要采用的方法做了简要介绍,阐述了各种方法的优缺点,同时介绍了字体库建立的算法。支持向量机是近几年来模式识别领域中的一个新技术,它被广泛应用到文字识别、人脸识别等应用中,是模式识别领域中的研究热点,本文提出了一种将支持向量机有效地用于解决多分类问题的策略。SVM具有结构简单,分类稳定可靠,且容错性好等优点。同时和LVQ 神经网络分类器识别方法进行了比较,计算机仿真表明,采用SVM用于手写体汉字识别更适合。最后根据目前的研究状况,指出今后研究中需要注意的问题和研究的发展方向。 关键词:汉字识别;预处理;特征提取;分类识别;支持向量机 Research on Recognition Methods of Chinese Handwriting Characters Abstract: Chinese handwriting characters recognition is the most difficult problem in character recognition field. It can improve nature and friendship when applied in natural human machine interactive areas, make a tremendous contribution to save manpower, improve efficiency and accelerate information transmission when applied in automatic text processing. This paper surveyed the main techniques in four phases: pre -processing, feature extracting, classification and recognition, and post-processing. The advantages and disadvantages of various methods were analyzed and the method of establishing font library was discussed. SVM (Support vector machine) is a new technique in pattern recognition in recent years, which is widely used in different fields such as characters and human face recognition and it has become a hotspot in pattern recognition. A strategy was proposed to solve handwriting digits multi--class classification problem effectively by SVM. SVM has merits of simple structure, steady and credible classification and good error tolerance. The simulation results show that SVM is more suitable for Chinese handwriting characters recognition compared with LVQ neural network classifier. In the end, some problems that should be paid more attention and some research development direction were proposed according to current research status. Keywords: Chinese character recognition, pre-processing, feature extraction, classification and recognition, support vector machine 第1章 绪 论 1
|
汉字手写体识别方法研究
查看评论
已有0位网友发表了看法