| 基于BP神经网络的数字识别方法 |
| |
| 引用本文： | 施少敏,马彦恒,陈建泗.基于BP神经网络的数字识别方法[J].兵工自动化,2006,25(10):40-41. |
| |
| 作者姓名： | 施少敏 马彦恒 陈建泗 |
| |
| 作者单位： | 军械工程学院,光学与电子工程系,河北,石家庄,050003;军械工程学院,光学与电子工程系,河北,石家庄,050003;军械工程学院,光学与电子工程系,河北,石家庄,050003 |
| |
| 摘 要： | 应用BP神经网络对数字进行识别,其图像的预处理采用去除杂点方法去除噪声,使用逐像素特征提取方法进行特征向量的提取.该识别过程还包括数字识别网络的训练,及通过选取初始权值和隐含层节点数、权值学习算法确定相关参数.实践证明,利用BP神经网络能够实现对阿拉伯数字的准确、快速识别.
|
| 关 键 词： | 神经网络 数字识别 BP算法 |
| 文章编号： | 1006-1576（2006）10-0040-02 |
| 收稿时间： | 2006-05-19 |
| 修稿时间： | 2006-07-16 |
Numerical Recognition Method Base on BP Neural Network |
| |
| SHI Shao-min,MA Yan-heng,CHEN Jian-si.Numerical Recognition Method Base on BP Neural Network[J].Ordnance Industry Automation,2006,25(10):40-41. |
| |
| Authors: | SHI Shao-min MA Yan-heng CHEN Jian-si |
| |
| Abstract: | Applied BP neural network to recognize numbers, and adopted a method of wipe off miscellaneous points to take out noise, and used a method of per pix feature extraction to extract feature vector. The course of recognition also includes training numerical recognition network and making sure correlation parameters by choosing the initial weight, the number of hide nodes and the learning algorithm of weight. Experiments indicated that the method work well in recognizing the Arab numbers exactly and quickly. |
| |
| Keywords: | Neural network Numerical recognition Back propagation algorithm |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
