| 利用汉明距离优选神经网络学习样本 |
| |
| 引用本文： | 申金媛,刘玥,张文伟,陈戍,郭鹏毅,宋庄,张延炘. 利用汉明距离优选神经网络学习样本[J]. 光学学报, 2000, 20(9): 1229-1234 |
| |
| 作者姓名： | 申金媛 刘玥 张文伟 陈戍 郭鹏毅 宋庄 张延炘 |
| |
| 作者单位： | 南开大学现代光学研究所,国家教育部光电信息科学技术开放实验室,天津,300071 |
| |
| 基金项目： | 国家自然科学基金!(6 98770 0 5)，国家科委 86 3高技术计划，天津市 2 1世纪青年基金，天津市高教局资助项目 |
| |
| 摘 要： | 鉴于学习样本对神经网络模型的模式识别性能有很大的影响 ,提出学习样本的选择应与识别模型所利用的特性相结合 ,并利用汉明 (Hamming)距离对用于旋转不变识别的级联模型的学习样本进行优选 ,计算机对三个很相似的飞机模型进行识别 ,识别结果表明对学习样本进行有效的选择不仅可以减少系统的学习训练时间而且可以提高模型的识别能力。
|
| 关 键 词： | 神经网络 模式识别 学习样本 级联模型 |
| 收稿时间： | 1999-01-15 |
Selecting of Learning Samples Based on Hamming Distance |
| |
| Shen Jinyuan,Liu Yue,Zhang Wenwei,Chen Shu,Guo Pengyi,Song Zhuang,Zhang Yanxin. Selecting of Learning Samples Based on Hamming Distance[J]. Acta Optica Sinica, 2000, 20(9): 1229-1234 |
| |
| Authors: | Shen Jinyuan Liu Yue Zhang Wenwei Chen Shu Guo Pengyi Song Zhuang Zhang Yanxin |
| |
| Abstract: | Learning samples significantly affect the recognition ability of neuron network models. One of selecting rules of learning samples is proposed according to the principle of the pattern recognition model. A method of selecting learning samples based on Hamming distance used in the cascade neuron network model for rotation invariance recognition is analyzed. The results of the computer recognition show that the effective selection of the learning samples can not only reduce the training time but also improve the recognition ability of the model. |
| |
| Keywords: | neural network pattern recognition learning sample the cascade neural network model. |
| 本文献已被 万方数据 等数据库收录！ |
|
