PCA和BP神经网络结合，适用于人脸识别中.zip

clear all clc close all %---------------------初始化------------------------ database='ORL';%使用的人脸库 rows=112; cols=92; ClassNum=40; tol_num=10; image_fmt='.bmp'; train_samplesize=5;%每类训练样本大小 address=[pwd,'\',database,'\s']; train_samp=1:train_samplesize;%训练样本 test_samp=train_samplesize+1:tol_num;%测试样本 [train_data,train_labels]=readsample(address,ClassNum,train_samp,rows,cols,image_fmt); [Pro_Matrix,Mean_Image]=my_pca(train_data,24); train_pro=Pro_Matrix'*train_data; train_T=ind2vec(train_labels);%训练样本期望输出 [pn,ps] = mapminmax(train_pro,-1,1); [tn,ts] = mapminmax(train_T,-1,1); net=newff(pn,tn,80,{'tansig','purelin'},'trainrp'); net.trainParam.epochs = 10000; net.trainParam.lr = 0.01; net.trainParam.goal = 1e-3; net.divideParam.trainRatio = 1; net.divideParam.valRatio = 0; net.divideParam.testRatio = 0; %网络训练 net=train(net,pn,tn); % 测试过程 [test_data,test_labels]=readsample(address,ClassNum,test_samp,rows,cols,image_fmt); test_pro=Pro_Matrix'*test_data; pnew=test_pro; pnewn = mapminmax('apply',pnew,ps); anewn = sim(net,pnewn); anew = mapminmax('reverse',anewn,ts); [out_max,index]=max(anew); accuracy=sum(test_labels==index)/length(test_labels)