| 一种新的BP神经网络人脸识别方法研究 |
| |
| 引用本文： | 范引娣,何正玲.一种新的BP神经网络人脸识别方法研究[J].西安文理学院学报,2007,10(1):93-96. |
| |
| 作者姓名： | 范引娣 何正玲 |
| |
| 作者单位： | 陕西交通职业技术学院电子商务教研室 陕西西安710014(范引娣)，西安文理学院图书馆 陕西西安710065(何正玲) |
| |
| 摘 要： | BP神经网络在人脸识别方面的研究中,原始样本数据不进行预处理与特征提取,不仅使识别结果准确降低,而且使BP神经的结构复杂化.主成分分析法能提取代替样本的少数几个主成分,这些主成分彼此不相关,符合特征优化的要求.在BioID人脸数据库实验表明,将主成分分析与BP神经网络相结合,与传统单一的BP神经网络识别相比,能够提高识别的正确率,减少训练时间,同时简化网络结构,减少很大的计算量.
|
| 关 键 词： | 主成分分析 BP神经网络 人脸识别 BioID人脸数据库 |
| 文章编号： | 1008-5564(2007)01-0093-04 |
| 修稿时间： | 2006年9月16日 |
A New Face Recognition Method of BP Neural Network |
| |
| FAN Yin-di,HE Zheng-ling.A New Face Recognition Method of BP Neural Network[J].Journal of Xi‘an University of Arts & Science:Natural Science Edition,2007,10(1):93-96. |
| |
| Authors: | FAN Yin-di HE Zheng-ling |
| |
| Abstract: | In the research of face recognition,based on BP neural network,the recognition precision will be low and the structure of BP neural network will become complex if samples' data are not preprocessed and the features are not extracted.The analysis principal components can extract a few main factors representing the whole samples,furthermore these factors are mutually uncorrelated,hence satisfying the requirements of features' optimization.BioID face database experiments show that compared with the traditional simple BP method,the combination of principal component analysis with BP neural network improves the precision,reduces the training time,simplifies the structure of network and considerably reduces the amount of calculation. |
| |
| Keywords: | analysis of principal components BP neural network face recognition BioID face database |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
