-
Notifications
You must be signed in to change notification settings - Fork 1
/
GenerateTrainingPatches_AISD_noSelect_label_multi.m
179 lines (140 loc) · 7.1 KB
/
GenerateTrainingPatches_AISD_noSelect_label_multi.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
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
%%% Generate the training data.
% 对分块后的数据进行筛选，剔除完全无阴影的块以及阴影像素个数较少的块
clear;close all;
addpath('utils');
batchSize = 10; %%% batch size
dataName = 'TrainingPatches';
folder_shadow = '..\datasets\AISD-Train412\shadow';
folder_mask = '..\datasets\AISD-Train412\mask';
patchsize = 256;
stride = 64;
step = 0;
count = 0;
ext = {'*.jpg','*.png','*.bmp','*.jpeg','*.tif'};
filepaths_shadow = [];
filepaths_mask = [];
for i = 1 : length(ext)
filepaths_shadow = cat(1,filepaths_shadow, dir(fullfile(folder_shadow, ext{i})));
filepaths_mask = cat(1,filepaths_mask, dir(fullfile(folder_mask, ext{i})));
end
%% count the number of extracted patches
% scales = [1 0.9 0.8 0.7];
scales = [1 0.8 0.6];
ImageNums = length(filepaths_mask);
for i = 1 : ImageNums
image_m = double(imread(fullfile(folder_mask,filepaths_mask(i).name)))/255;
image_m = image_m(:,:,1);
if mod(i,100)==0
disp([i,length(filepaths_mask)]);
end
for s = 1:1
image_m_scale = imresize(image_m,scales(s),'bicubic');
[hei,wid,~] = size(image_m_scale);
for x = 1+step : stride : (hei-patchsize+1)
for y = 1+step :stride : (wid-patchsize+1)
mask_patch = image_m_scale(x : x+patchsize-1, y : y+patchsize-1,:);
for j = 1:1
mask_aug = data_augmentation(mask_patch, j);
count = count+1;
% % 判断掩膜块中是否有阴影区域，且阴影像素个数大于50；如满足条件则加以统计
% if(sum(mask_aug(:))>50)
% count = count+1;
% end
end
end
end
end
end
numPatches = ceil(count/batchSize)*batchSize;
disp('TotalPatchNumbers batchSize Iterations');
disp([numPatches,batchSize,numPatches/batchSize]);
%pause;
%% extract patches
inputs_shadow = zeros(patchsize, patchsize, 3, numPatches,'single'); % this is fast
% inputs_ratio = zeros(patchsize, patchsize, 1, numPatches,'single'); % this is fast
% inputs_ratio2 = zeros(patchsize/2, patchsize/2, 1, numPatches,'single'); % this is fast
% inputs_ratio4 = zeros(patchsize/4, patchsize/4, 1, numPatches,'single'); % this is fast
% inputs_ratio8 = zeros(patchsize/8, patchsize/8, 1, numPatches,'single'); % this is fast
inputs_label = zeros(patchsize, patchsize, 1, numPatches,'single'); % this is fast
inputs_label2 = zeros(patchsize/2, patchsize/2, 1, numPatches,'single'); % this is fast
inputs_label4 = zeros(patchsize/4, patchsize/4, 1, numPatches,'single'); % this is fast
inputs_label8 = zeros(patchsize/8, patchsize/8, 1, numPatches,'single'); % this is fast
count = 0;
tic;
for i = 1 : ImageNums
image_s = double(imread(fullfile(folder_shadow,filepaths_shadow(i).name)))/255; %
image_m = double(imread(fullfile(folder_mask,filepaths_mask(i).name))); %
image_m = image_m(:,:,1);
image_m(image_m>0)=1;
image_m(image_m<0)=0;
image_m = image_m+1;
% %Compute (H +1)/(I +1) ratio images
% hsi = rgb2hsi(image_s);
% hsi_h = hsi(:,:,1);
% hsi_s = hsi(:,:,2);
% hsi_i = hsi(:,:,3);
% ratio = (hsi_h + 1)./(hsi_i + 1);
%
% %Normalized (H +1)/(I +1) ratio images
% min1 = min(ratio(:));
% max1 = max(ratio(:));
% ratio = (ratio - min1)/(max1 - min1);
if mod(i,100)==0
disp([i,length(filepaths_shadow)]);
end
for s = 1:1
image_s_scale = im2single(imresize(image_s,scales(s),'bicubic'));
image_m_scale = im2single(imresize(image_m,scales(s),'nearest'));
% image_r_scale = im2single(imresize(ratio,scales(s),'bicubic'));
[hei,wid,bands] = size(image_s_scale);
%
% % 合并比例图到输入影像中
% image_s_scale(:,:,bands+1) = image_r_scale;
for x = 1+step : stride : (hei-patchsize+1)
for y = 1+step :stride : (wid-patchsize+1)
image_patch = image_s_scale(x : x+patchsize-1, y : y+patchsize-1,:);
% ratio_patch = image_r_scale(x : x+patchsize-1, y : y+patchsize-1,:);
mask_patch = image_m_scale(x : x+patchsize-1, y : y+patchsize-1,:);
for j = 1:1
image_aug = data_augmentation(image_patch, j); % augment data
% ratio_aug = data_augmentation(ratio_patch, j); % augment data
mask_aug = data_augmentation(mask_patch, j); % augment data
count = count+1;
inputs_shadow(:, :, :, count) = image_aug;
% % Multi-ratio
% inputs_ratio(:, :, :, count) = ratio_aug;
% inputs_ratio2(:, :, :, count) = imresize(ratio_aug, 0.5);
% inputs_ratio4(:, :, :, count) = imresize(ratio_aug, 0.25);
% inputs_ratio8(:, :, :, count) = imresize(ratio_aug, 0.125);
% Multi-label
inputs_label(:, :, :, count) = mask_aug;
inputs_label2(:, :, :, count) = imresize(mask_aug, 0.5, 'nearest');
inputs_label4(:, :, :, count) = imresize(mask_aug, 0.25, 'nearest');
inputs_label8(:, :, :, count) = imresize(mask_aug, 0.125, 'nearest');
% % 判断掩膜块中是否有阴影区域，且阴影像素个数大于50；如满足条件则加以统计
% if(sum(mask_aug(:))>50)
% count = count+1;
% inputs_shadow(:, :, :, count) = image_aug;
% inputs_label(:, :, :, count) = mask_aug;
% end
end
end
end
end
end
toc;
set = uint8(ones(1,size(inputs_shadow,4)));
disp('-------Datasize-------')
disp([size(inputs_shadow,4),batchSize,size(inputs_shadow,4)/batchSize]);
if ~exist(dataName,'file')
mkdir(dataName);
end
%% save data
tic;
%save(fullfile(dataName,['imdb_rgb_',num2str(patchsize),'P_',num2str(batchSize),'B_',num2str(ImageNums),'_aug_select']), 'inputs_shadow','inputs_label','set','-v7.3');
% save(fullfile(dataName,['imdb_rgb_',num2str(patchsize),'P_AISD',num2str(ImageNums),'_noSelect_ratio_label_multi']), ...
% 'inputs_shadow','inputs_ratio','inputs_ratio2','inputs_ratio4','inputs_ratio8',...
% 'inputs_label','inputs_label2','inputs_label4','inputs_label8', 'set','-v7.3');
save(fullfile(dataName,['imdb_rgb_',num2str(patchsize),'P_AISD',num2str(ImageNums),'_noSelect_label_multi']), ...
'inputs_shadow', 'inputs_label','inputs_label2','inputs_label4','inputs_label8', 'set','-v7.3');
toc;