Repository navigation
Expand file tree
/
Copy pathmain.m
More file actions
126 lines (117 loc) · 4.42 KB
/
Copy pathmain.m
File metadata and controls
126 lines (117 loc) · 4.42 KB
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
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
function [] = main()
% - FINAL PROJECT -
%
% FACE DETECTION USING VIOLA-JONES DETECTOR
%
% Nenad Mancevic, July 2012.
% The Faculty of Electronic Engineering Nis (c) 2012.
%-------------------------------------------------------------------------
clear all
addpath('I:\Work\Face Detection\Project\GML_AdaBoost_Matlab_Toolbox_0.3');
% some initialization values
posImages = 150;
negImages = 200;
totalImages = 350;
totalNoOfFeatures = 35802;
positiveFeatures = zeros(posImages, totalNoOfFeatures);
negativeFeatures = zeros(negImages, totalNoOfFeatures);
% 1. Load training set
% if it's not precomputed, precompute it and save it for later use
if (exist('posFeatures.mat', 'file') == 0)
for k=1:posImages
if (k < 10)
str = strcat('mit\train\face\face0000', int2str(k), '.pgm');
elseif (k < 100)
str = strcat('mit\train\face\face000', int2str(k), '.pgm');
elseif (k < 1000)
str = strcat('mit\train\face\face00', int2str(k), '.pgm');
elseif (k < 10000)
str = strcat('mit\train\face\face0', int2str(k), '.pgm');
end
currImg = imread(str);
if (mod(k,100)==0)
fprintf('Extracting features from positive image: %d\n', k);
end
% 2. Extract features out of it
positiveFeatures(k, :) = extractFeatures(currImg);
% we can turn on mode in extractFeatures to extract all features' info
% in terms of x,y position, size, type, and so on - useful for
% detection later
% save featuresInfo.mat featuresInfo
end
save posFeaturesBig.mat positiveFeatures
else
positiveFeatures = load('posFeatures.mat');
end
% same for negative features
if (exist('negFeatures.mat', 'file') == 0)
j=1;
while j <= negImages
if (j < 10)
str = strcat('mit\train\non-face\B1_0000', int2str(j), '.pgm');
elseif (j < 100)
str = strcat('mit\train\non-face\B1_000', int2str(j), '.pgm');
elseif (j < 1000)
str = strcat('mit\train\non-face\B1_00', int2str(j), '.pgm');
elseif (j < 10000)
str = strcat('mit\train\non-face\B1_0', int2str(j), '.pgm');
end
if (exist(str))
currImg = imread(str);
if (mod(j,100)==0)
fprintf('Extracting features from negative image: %d\n', j);
end
negativeFeatures(k, :) = extractFeatures(currImg);
j=j+1;
end
end
save negFeaturesBig.mat negativeFeatures
else
negativeFeatures = load('negFeatures.mat');
end
% 3. Train the classifiers using the features
weak_learner = tree_node_w(3); % our weak learners
all_features = [positiveFeatures.positiveFeatures; negativeFeatures.negativeFeatures];
labels(1:posImages) = 1;
labels(posImages+1:totalImages) = -1;
all_features = all_features';
% learning algorithm wants Features X NoImages instead od ImagesXFeatures
if (exist('learners.mat')==0 & exist('weights.mat')==0)
fprintf('Training starts...');
c=clock;
fprintf('Current time: %d:%d:%1.f\n', c(4), c(5), c(6));
[learners weights] = RealAdaBoost(weak_learner, all_features, labels, 100);
save learnersBig.mat learners
save weightsBig.mat weights
fprintf('Training finishes...');
e=clock;
fprintf('Current time: %d:%d:%1.f\n', e(4), e(5), e(6));
else
fprintf('Classifiers already trained. Loading...');
learners = load('learners.mat');
weights = load('weights.mat');
end
% 4. Load testing set
testimg = imread('testImgs\profile.jpg');
testimg2 = rgb2gray(testimg); % convert to grayscale
resized = imresize(testimg2,1/8); % resize it for faster processing
n = size(resized, 1);
m = size(resized, 2);
% 5. Process images, using learned classifiers
imshow(testimg);
count = 0;
for y=2:1:m-18
for x=2:1:y
% x
% y
count = count+1;
subimg = resized(x:x+19-1,y:y+19-1);
if (detectFace(learners.learners, weights.weights, subimg) == 1)
% winFeatures = extractFeatures(subimg);
% if (Classify(learners.learners, weights.weights, winFeatures) > 1) %(detectFace(learners.learners, weights.weights, winFeatures) == 1)
rectangle('Position', 8*[y x 19 19], 'EdgeColor', 'g');
end
end
end
fprintf('Total number of iterations: %d\n', count);
end