一种新的降维算法PCA_LLE在图像识别中的应用  

Application of a new dimensionality reduction algorithm PCA_LLE in image recognition

作　　者：蓝雯飞[1] 汪敦志 张盛兰 LAN Wenfei;WANG Dunzhi;ZHANG Shenglan(College of Computer Science,South⁃Central University for Nationalities,Wuhan 430074,China)

机构地区：[1]中南民族大学计算机科学学院

出　　处：《中南民族大学学报（自然科学版）》2020年第1期85-90,共6页Journal of South-Central University for Nationalities：Natural Science Edition

基　　金：国家自然科学基金资助项目(61379059)

摘　　要：通过结合PCA与LLE两种降维方法,提出新的PCA_LLE算法,使它们优势互补.在手写体数字数据集上进行实验,先对数据集降维,再用K近邻算法对降维后的数据分类.实验结果表明融合两种算法的PCA_LLE降维方法较原来的PCA和LLE算法准确率均有了提升.而且新算法PCA_LLE对新样本的降维时间较LLE算法减少很多.在ORL人脸数据集上的实验表明,PCA_LLE算法较PCA,LLE算法准确率有所提高.Combining two dimensionality reduction methods(PCA and LLE),we propose a new PCA_LLE algorithm to absorb each other′s advantages.We conduct an experiment on handwritten digital dataset,first reduce the dimension of the dataset, and then use the K-nearest neighbor algorithm to classify the dimensionality-reduced data.Experiment results show that the accuracy of the dimensionality reduction methods combining the two algorithms are improved compared with the original PCA and LLE algorithms.Moreover,the new algorithm PCA_LLE reduces the dimension reduction time of the new sample much more than the LLE algorithm.Experiments on the ORL face database show that the PCA_LLE algorithm has higher accuracy than the PCA and LLE algorithms.

关 键 词：降维 图像识别 PCA_LLE算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]