-
Notifications
You must be signed in to change notification settings - Fork 12
/
ExtractAndTrain.m
104 lines (80 loc) · 2.97 KB
/
ExtractAndTrain.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
%崔佳勋 2017-5-20
clear
clc
%开始提取Haar-Like 特征
positiveRange=1:1000;
negativeRange=1:2000;
positiveNum=length(positiveRange); % 正样本数量
negativeNum=length(negativeRange); % 负样本数量
n=negativeNum+positiveNum;
stdSize=[20 20]; %MIT人脸库素材大小
HarrLike{1}=[1 -1]; %定义5种haar-like特征
HarrLike{2}=[1 -1].';
HarrLike{3}=[1 -2 1];
HarrLike{4}=[1 -2 1].';
HarrLike{5}=[1 -1;-1 1];
baseSize=2:2:4;
features=zeros(n,78460);
for i=positiveRange
picName=strcat('face',num2str(i),'.bmp'); %自动生成图片名,对图片进行遍历
%picName='face1.bmp';
FaceRegion=imread(picName);
if numel(size(FaceRegion))>2
FaceRegion=rgb2gray(FaceRegion); %将rgb图像转化为灰度图像
end
FaceRegion=imresize(FaceRegion,stdSize);
[II]=integralImage(FaceRegion);
[fea]=extHarrLikeFeature(II,HarrLike,baseSize);
features(i,:)=fea';
disp(picName);
end
for i=negativeRange
picName=strcat('nonface',num2str(i),'.bmp');
FaceRegion=imread(picName);
if numel(size(FaceRegion))>2
FaceRegion=rgb2gray(FaceRegion); %将rgb图像转化为灰度图像
end
FaceRegion=imresize(FaceRegion,stdSize);
[II]=integralImage(FaceRegion);
[fea]=extHarrLikeFeature(II,HarrLike,baseSize);
features(i+positiveNum,:)=fea';
disp(picName);
end
Y=[ones(1,positiveNum) zeros(1,negativeNum)];% 类标
HarrLikeFeatures=features;
save HarrLikeFeatures-2.mat HarrLikeFeatures Y
%}
%开始训练
trainingRate=0.5; % 训练样本占总体样本比例
T=200; % 训练轮数的范围
%T=100; % 训练轮数的范围
%
% trainError=zeros(1,T); % 训练错误率矩阵
% testError=zeros(1,T); % 训练错误率矩阵
% testTP=zeros(1,T); % TRUE-Positive 比例矩阵
% testFP=zeros(1,T); % FALSE-Positive 比例矩阵
% AdaCostTime=cell(1); % 花费时间
rows=size(features,1);
% 确实训练与测试样本
trainingRows=1:2:rows;
testingRows=2:2:rows;
trainX=features(trainingRows,:); % 获得训练数据
trainY=Y(trainingRows); % 训练数据的类标
testX=features(testingRows,:); % 获得测试数据
testY=Y(testingRows); % 测试数据的类标
[Hypothesis,AlphaT,trainError,AdaCostTime]=trainAdaBoostLearner(trainX,trainY,T);
[testError,testTP,testFP]=testAdaBoostLearner(testX,testY,Hypothesis,AlphaT,T);
% 下面一部分是作图的程序
% 该图形显示 获得训练错误率和测试错误率
figure(1002);hold on,
grid on,
xlabel('训练轮数');
ylabel('AdaBoost分类器错误率) ');
testingNum=ceil((1-trainingRate)*size(features,1)); %测试样本数量
trainingNum=size(features,1)-testingNum; %训练样本数量
title(strcat('AdaBoost训练与测试错误率',' ( 训练',num2str(trainingNum),'例,测试',num2str(testingNum),'例 )'));
testRange=1:T;
plot(testRange,mean(trainError,1),'r-');
plot(testRange,mean(testError,1),'b-');
legend('Harr-Like特征训练错误率','Harr-Like特征测试错误率');
save HarrLikeFeatures-2.mat HarrLikeFeatures Y trainError testError testTP testFP Hypothesis AlphaT AdaCostTime