CLBP人脸识别程序及运行结果

clear all;%只做M的特征的识别 load AR_label load ar_2;%可变换库 %///////////////////////////// %修改参数区 sample=59; cut_num=3; R=1; P=8; MAPPING=getmapping(P,'u2'); %//////////////////////////// class_num = 119; sample_num = 26; right_num = []; row=60; line=60; for test_mode = 6 for loop = 1 train_num = test_mode;%每类的训练样本是6个 train_total = train_num * class_num;%所有类得训练样本总数6*26=156 test_num = sample_num - train_num;%测试样本的个数 test_total = test_num * class_num;%所有类得测试样本总数 ar_train = zeros(row,line, train_total);%0矩阵的大小行数为每类的行*列，列为所有的训练样本的个数 ar_test = zeros(row,line, test_total); for i = 1:class_num kk1 = (i-1)*sample_num + label(loop, 1:train_num); kk2 = (i-1)*train_num + 1 : i*train_num; ar_train(:,:,kk2) = ar_2(:,:,kk1); kk1 = (i-1)*sample_num + label(loop, train_num+1:sample_num); kk2 = (i-1)*test_num + 1 : i*test_num; ar_test(:,:,kk2) = ar_2(:,:,kk1); end end end for i=1:train_total ar_train(:,:,i)=gaosi(ar_train(:,:,i));%预处理 end for i=1:test_total ar_test(:,:,i)=gaosi(ar_test(:,:,i)); end ar_train_new = zeros(row/cut_num,line/cut_num, cut_num*cut_num,train_total); ar_test_new = zeros(row/cut_num,line/cut_num, cut_num*cut_num,test_total); ar_train_new=reshape_1(ar_train,cut_num); ar_test_new=reshape_1(ar_test,cut_num); train_new_q=zeros(sample,train_total*cut_num*cut_num,2); test_new_q=zeros(sample,test_total*cut_num*cut_num,2); for i=1:train_total for j=1:cut_num*cut_num [S,M,C]=clbp(ar_train_new(:,:,j,i),R,P,MAPPING,'hist'); train_new_q(:,(i-1)*cut_num*cut_num+j,1)=reshape(S,[sample,1]); train_new_q(:,(i-1)*cut_num*cut_num+j,2)=reshape(M,[sample,1]); end end for i=1:test_total for j=1:cut_num*cut_num [S,M,C]=clbp(ar_test_new(:,:,j,i),R,P,MAPPING,'hist'); test_new_q(:,(i-1)*cut_num*cut_num+j,1)=reshape(S,[sample,1]); test_new_q(:,(i-1)*cut_num*cut_num+j,2)=reshape(M,[sample,1]); end end % train_new_S=zeros(cut_num*cut_num*sample,train_total); % test_new_S=zeros(cut_num*cut_num*sample,test_total); % % train_new_M=zeros(cut_num*cut_num*sample,train_total); % test_new_M=zeros(cut_num*cut_num*sample,test_total); % % train_new=zeros(cut_num*cut_num*sample*2,train_total); % test_new=zeros(cut_num*cut_num*sample*2,test_total); train_new_S=reshape_2(train_new_q(:,:,1),cut_num); test_new_S=reshape_2(test_new_q(:,:,1),cut_num); train_new_M=reshape_2(train_new_q(:,:,2),cut_num); test_new_M=reshape_2(test_new_q(:,:,2),cut_num); train_new=[train_new_S;train_new_M]; test_new=[test_new_S;test_new_M]; error = []; for i = 1:class_num for j = 1:test_num %i,j kk = (i-1)*test_num + j; res_y = test_new(:,kk); dis = []; for ii = 1:class_num for jj = 1:train_num res_x = train_new(:,(ii-1)*train_num+jj); res = (res_y - res_x).^2; temp = sqrt(sum(res)); dis = [dis; temp]; end end [value, site] = min(dis); site_new = floor((site-1)/train_num) + 1; if site_new ~= i error = [error; i*test_num+j]; end end end right_num = [right_num; (1.0-length(error)/(class_num*test_num))*100 ]; right_num'