| 一种基于RBF神经网络的英文字符识别方法 |
| |
| 引用本文： | 邱望成,刘春林,岳振军. 一种基于RBF神经网络的英文字符识别方法[J]. 解放军理工大学学报(自然科学版), 2003, 4(6): 33-36 |
| |
| 作者姓名： | 邱望成 刘春林 岳振军 |
| |
| 作者单位： | [1]解放军理工大学理学院，江苏南京211101 [2]解放军理工大学通信工程学院，江苏南京210007 |
| |
| 摘 要： | 提出了一种基于RBF神经网络的英文字符识别方法。该方法首先提取字符的结构特征和统计特征，以它们作为神经网络的输入向量，然后用RBF神经网络进行识别。使用了高斯函数作为神经网络的激励函数，并以最小二乘准则对字符进行识别。对字符样本的识别结果显示，此方法在识别错误率和识别效率等方面均有很好的效果。
|
| 关 键 词： | 字符识别 统计特征 结构特征 RBF神经网络 |
| 文章编号： | 1009-3443(2003)06-0033-04 |
| 修稿时间： | 2003-03-24 |
New Approach to English Character RecognitionBased on RBF Neural Network |
| |
| QIU Wang-cheng,LIU Chun-lin and YUE Zhen-jun. New Approach to English Character RecognitionBased on RBF Neural Network[J]. Journal of PLA University of Science and Technology(Natural Science Edition), 2003, 4(6): 33-36 |
| |
| Authors: | QIU Wang-cheng LIU Chun-lin YUE Zhen-jun |
| |
| Affiliation: | QIU Wang-cheng~1,LIU Chun-lin~2,YUE Zhen-jun~2 |
| |
| Abstract: | A new approach to English character recognition based on RBF neural network is presented . By this approach, the structural features and statistical features of characters are firstly extracted, then they are normalized to the input vector of neural network, and RBF neural network is used for recognition. Gauss function is used as neural network's inspirit function , and least square rule is used to recognize the character. Recognition result based on the character sample shows that the method achieves excellent performance in terms of recognition error rates and recognition efficiency. |
| |
| Keywords: | character recognition statistical characteristic structural characteristic RBF neural network |
| 本文献已被 CNKI 维普 等数据库收录！ |
| 点击此处可从《解放军理工大学学报(自然科学版)》浏览原始摘要信息 |
|
点击此处可从《解放军理工大学学报(自然科学版)》下载免费的PDF全文 |
