毕业论文（设计）基于人工神经网络的手写识别系统.doc

摘　　要 信息技术的快速发展，计算机迅速走进人们的生活，手写识别技术 本文中利用神经网络算法识别数字图像和手写汉字。针对数字图像首先特征矩阵然后用神经网路算法识别，进而获得识别的结果。针对汉字识别，首先处理图像获得特征矩阵，然后用笔画去识别汉字，最后获得识别的汉字。 关键词：手写识别；图像处理；神经 ABSTRACT The rapid development of information technology makes the computer immediately appear in peoples lives. The application of handwriting recognition technology becomes widespread, and there are more and more application areas. For example, we can see it on a mobile deviceand in the printing work. It makes peoples daily life, work and learning more convenient. Therefore, the handwriting recognition has very strong applied value and practical value. In this article, the writer uses neural network to identify digital image and handwriting Chinese ideogram. The premise of recognition is to obtain a stable neural network and a great deal of sample training is needed. For digital image, we need to process the images, and get characteristic matrix through a series of algorithms, including graying, linearization, Median filter, gradient sharpening and normalization.For Chinese ideogram recognition,we should gain a characteristic matrix by processing images, and then get painting pen through the training neural network identification.Next, we use painting pen to recognize Chinese ideogram.At last, the Chinese ideogram and associative Chinese characters and phrases will be obtained. After testing and verification, handwriting recognition which is based on artificial neural network system can effectively identify the handwriting image and Chinese ideogram. We obtain satisfying results in the identification test on words and digital image. Key words: Handwriting recognition;image processing;neural network 目　　录 第1章 绪论 1 1.1 课题研究的背景 1 1.2 课题研究的目的及意义 1 1.3 国内外研究现状 2 1.4 课题研究内容 2 1.5 论文的组织结构 3 第2章 神经网络算法的原理 4 2.1 神经网络 4 2.1.1 生物神经元网络 4 2.1.2 人工神经元网络 4 2.2 神经元学习算法 6 2.2.1 前馈神经网络 6 2.2.2 感知机 6 2.2.3 反向传播算法 8 2.2.4 神经网络在模式识别上面的优势 11 2.3 本章小结 11 第3章 系统设计与实现 12 3.1 神经网络算法的实现 12 3.1.1 神经网络的结构 12 3.1.2 算法的结构 12 3.2 神经网络识别手写数字 14 3.