分析了基于总体离散度矩阵和总类间离散度矩阵的主成分分析的原理。利用两种方法分别提取人脸特征并进行识别。对两种方法获得的结果进行了特征层融合和决策层融合,基于ORL人脸数据库的实验表明该方法的识别性能优于单一的主成分分析方法。
The principle of two different PCAs,PCA based on global scatter matrix and PCA based on global between-class scatter matrix is analyzed firstly.Two different fusion methods,feature level fusion and decision level fusion are proposed using the feature got from two different PCAs.The experiment result is displayed and data fusion method is proved to be efficient for getting better recognition rate.