规则模糊系统和ICA在图像识别中的应用

Fuzzy Rule-based System and Application of ICA in Image Recognition

利用模糊系统中的模糊规则和SVM中内积函数之间的关系,通过支持向量机对训练样本进行学习,由支持向量确定模糊规则,构建基于规则的模糊分类器。对人脸图像做独立成分分析,提取有利于分类的面部特征的主要独立成分。该模糊系统在ORL人脸库上的实验结果表明,该方法同样适合解决小样本、非线性、高维模式的分类问题。

Exploiting the relationships between fuzzy rules of fuzzy system and inner product function of SVM,learning the training samples with SVM,a fuzzy classifier can be built based on fuzzy rules given by SVM.Furthermore,the ICA of face image is used to get the main independent component of facial features. Based on ORL face database,experimental results of this fuzzy system show that this method is also applicable to solve the classification problems of small samples,nonlinearity and high dimensions.