摘要
提出了一种新的基于独立成分分析和径向基神经网络的人脸识别方法。独立成分分析可以从高阶上消除特征数据的相关性,改进了主成分分析方法只能从2阶上消除数据相关性的弱点;最终特征数据的分类由RBF神经网络来实现。在人脸数据库上的实验结果表明该新方法的识别性能较其他方法有了很大提高。
A new method of face recognition based on the independent component analysis and the neural network with radial basic Function is proposed. The independent component analysis can eliminate the data correlation of a higher order while the principal component analysis can only eliminate the data correlation of the second order. The original samples after feature extraction are classified by neural network with radial basic function. Experimental results on the face database demonstrate the efficiency of the new algorithm.
出处
《江苏科技大学学报（自然科学版）》
CAS
北大核心
2006年第4期46-50,共5页
Journal of Jiangsu University of Science and Technology:Natural Science Edition
关键词
人脸识别
特征提取
独立成分分析
主成分分析
径向基神经网络
face recognition
feature extraction
independent component analysis
principal component analysis
neural network with radial basic function
