-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathtolerance_test_ICIP.m
269 lines (218 loc) · 8.24 KB
/
tolerance_test_ICIP.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
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
function [ goodm, badm ] = tolerance_test_ICIP( im1, im2, im3, patch_dim, N_rep, t, phi )
%GENERATE_TEST Summary of this function goes here
% Detailed explanation goes here
DESCVEC = 1:7;
descname = {'SIFT' 'RootSIFT' 'Rafa_weights' 'Rafa_quantised' 'SURF' 'SIFT_0.6' 'RootSIFT_0.6'};
goodm = zeros(1,max(DESCVEC));
badm = zeros(1,max(DESCVEC));
LIBRARYPATH = 'LD_LIBRARY_PATH= ';
MATCHER_PATH_global = './siim-matcher';
VARIANCE = 0.005;
SHOWON = false;
PERFORM_IMAS_MEX = true;
DO_MOD_CHECK = true;
im1 = sum(im1,3)/3; im1 = im1*( 255/max(im1(:)) );
im2 = sum(im2,3)/3; im2 = im2*( 255/max(im2(:)) );
im3 = sum(im3,3)/3; im3 = im3*( 255/max(im3(:)) );
[x_im1,y_im1] = size(im1);
[x_im2,y_im2] = size(im2);
[x_im3,y_im3] = size(im3);
x = randi([1 x_im3-patch_dim]);
y = randi([1 y_im3-patch_dim]);
patch = im3(x:x+patch_dim-1, y:y+patch_dim-1);
patch = repmat(patch, N_rep);
%imshow(patch/max(patch(:)));
% image 1
x1 = randi([1 x_im1-N_rep*patch_dim]);
y1 = randi([1 y_im1-N_rep*patch_dim]);
x1_top = x1+N_rep*patch_dim-1;
y1_top = y1+N_rep*patch_dim-1;
%figure;
im1(x1:x1_top,y1:y1_top) = patch;
if (VARIANCE>0)
im1 = imnoise(im1/max(im1(:)),'gaussian',0,VARIANCE);
im1 = im1*( 255/max(im1(:)) );
end
%imshow(im1/max(im1(:)));
% image 2
x2 = randi([1 x_im2-N_rep*patch_dim]);
y2 = randi([1 y_im2-N_rep*patch_dim]);
x2_top = x2+N_rep*patch_dim-1;
y2_top = y2+N_rep*patch_dim-1;
%figure;
im2(x2:x2_top,y2:y2_top) = patch;
im2 = perform_tilt_on_image(im2,t,phi);
if (VARIANCE>0)
im2 = imnoise(im2/max(im2(:)),'gaussian',0,VARIANCE);
im2 = im2*( 255/max(im2(:)) );
end
%imshow(im2/max(im2(:)));
for DESC=DESCVEC
opts_IMAS.uno = 1;
data_matches = [];
MATCHER_PATH = [MATCHER_PATH_global '/' descname{DESC}];
mkdir(MATCHER_PATH);
num_matches = 0;
if(~PERFORM_IMAS_MEX)
imwrite(im1/max(im1(:)),[MATCHER_PATH '/im1.png']);
imwrite(im2/max(im2(:)),[MATCHER_PATH '/im2.png']);
delete([MATCHER_PATH '/data_matches.csv']);
switch DESC
case 1 % SIFT L1 0.8
[~] = system([LIBRARYPATH ' cd ' MATCHER_PATH ' && ' LIBRARYPATH ' ../main -im1 im1.png -im2 im2.png -applyfilter 0 -match_type 0 -covering 1.7 -desc 1 -match_ratio 0.8' ]);
case 2 % RootSIFT 0.8
[~] = system([LIBRARYPATH ' cd ' MATCHER_PATH ' && ' LIBRARYPATH ' ../main -im1 im1.png -im2 im2.png -applyfilter 0 -match_type 0 -covering 1.9 -desc 11 -match_ratio 0.8' ]);
case 6 % SIFT L1 0.6
[~] = system([LIBRARYPATH ' cd ' MATCHER_PATH ' && ' LIBRARYPATH ' ../main -im1 im1.png -im2 im2.png -applyfilter 0 -match_type 0 -covering 1.7 -desc 1 -match_ratio 0.6' ]);
case 7 % RootSIFT 0.6
[~] = system([LIBRARYPATH ' cd ' MATCHER_PATH ' && ' LIBRARYPATH ' ../main -im1 im1.png -im2 im2.png -applyfilter 0 -match_type 0 -covering 1.9 -desc 11 -match_ratio 0.6' ]);
case 3 % Rafa weights
[~] = system([LIBRARYPATH ' cd ' MATCHER_PATH ' && ' LIBRARYPATH ' ../main -im1 im1.png -im2 im2.png -applyfilter 0 -match_type 1 -covering 1.6 -AC_type 1' ]);
case 4 % Rafa quantised
[~] = system([LIBRARYPATH ' cd ' MATCHER_PATH ' && ' LIBRARYPATH ' ../main -im1 im1.png -im2 im2.png -applyfilter 0 -match_type 1 -covering 1.7 -AC_type 2 -new_ori_size 22 -step_sigma 1.5 -quant_prec 0.32' ]);
case 5 % SURF
%[~] = system([LIBRARYPATH ' cd ' MATCHER_PATH ' && ' LIBRARYPATH ' ../main -im1 im1.png -im2 im2.png -applyfilter 0 -match_type 0 -covering 1 -desc 2 -match_ratio 0.6' ]);
end
[~,cmdout] = system([LIBRARYPATH 'cat ' MATCHER_PATH '/data_matches.csv | wc -l' ]);
if (exist([MATCHER_PATH '/data_matches.csv'], 'file') == 2)
num_matches = str2num(cmdout)-1;
end
else
opts_IMAS.applyfilter = 0;
opts_IMAS.showon = false;
switch DESC
case 1 % SIFT L1 0.8
opts_IMAS.covering = 1.7;
opts_IMAS.desc_type = 1;
opts_IMAS.match_ratio = 0.8;
case 2 % RootSIFT 0.8
opts_IMAS.covering = 1.9;
opts_IMAS.desc_type = 11;
opts_IMAS.match_ratio = 0.8;
case 6 % SIFT L1 0.6
opts_IMAS.covering = 1.7;
opts_IMAS.desc_type = 1;
opts_IMAS.match_ratio = 0.6;
case 7 % RootSIFT 0.6
opts_IMAS.covering = 1.9;
opts_IMAS.desc_type = 11;
opts_IMAS.match_ratio = 0.6;
case 3 % Rafa weights
opts_IMAS.covering = 1.6;
opts_IMAS.desc_type = 31;
case 4 % Rafa quantised
opts_IMAS.covering = 1.7;
opts_IMAS.desc_type = 32;
case 5 % SURF
opts_IMAS.covering = 1.7;
opts_IMAS.desc_type = 2;
end
opts_IMAS.covering = 1;
if (opts_IMAS.desc_type > 0)
data_matches = perform_IMAS(im1, im2,opts_IMAS);
num_matches = size(data_matches,2);
else
num_matches = 0;
end
end
if (num_matches>0)
if(~PERFORM_IMAS_MEX)
data_matches = csvread([MATCHER_PATH '/data_matches.csv'],1,0)';
end
data_x1 = data_matches(2,:);
data_y1 = data_matches(1,:);
data_x2 = data_matches(9,:);
data_y2 = data_matches(8,:);
x1p=0; y1p=0; x2p=0 ;y2p =0;
good = true(1,length(data_x1));
% change of coordinates x <-> y
for i=1:length(data_x1)
x0 = data_x1(i);
y0 = data_y1(i);
if ( (x1<=x0)&&(x0<=x1_top)&&(y1<=y0)&&(y0<=y1_top) )
x1p = mod(x0-x1,patch_dim);
y1p = mod(y0-y1,patch_dim);
good(i) = true*good(i);
else
good(i) = false*good(i);
end
[y0,x0,inside] = get_tilted_coor( data_y2(i), data_x2(i), y_im2, x_im2, t, phi*pi/180, -1);
if (inside)
if ( (x2<=x0)&&(x0<=x2_top)&&(y2<=y0)&&(y0<=y2_top) )
if (DO_MOD_CHECK)
x2p = mod(x0-x2,patch_dim);
y2p = mod(y0-y2,patch_dim);
if ( (norm([x1p-x2p, y1p-y2p])<=10) )
good(i) = true*good(i);
else
good(i) = false*good(i);
end
else
good(i) = true*good(i);
end
else
good(i) = false*good(i);
end
else
good(i) = false*good(i);
end
end
goodm(DESC) = sum(good);
badm(DESC) = sum(1-good);
if (SHOWON)
pause(0.2);
figure;
plot_matches_in_images(im1,im2,data_matches(:,good==true));
title([num2str(sum(good)) ' good matches for ']);
pause(0.2);
figure;
plot_matches_in_images(im1,im2,data_matches(:,good==false));
title([num2str(sum(1-good)) ' false matches for ']);
end
end
end
%
% %im1
% imseg = zeros(size(im1));
% for x=1:10:size(im1,1)
% for y=1:10:size(im1,2)
% % change of coordinates x <-> y
% x0 = x;
% y0 = y;
%
% if ( (x1<=x0)&&(x0<=x1_top)&&(y1<=y0)&&(y0<=y1_top) )
% imseg(x,y) = 1; % good matches / target region
% else
% imseg(x,y) = 0.5;
% end
%
% end
% end
% figure;
% imshow(imseg);
%
%
%
%
%
%
%
% %im2
% imseg = zeros(size(im2));
% for x=1:5:size(im2,1)
% for y=1:5:size(im2,2)
% % change of coordinates x <-> y
% [y0,x0,inside] = get_tilted_coor( y, x, y_im2, x_im2, t, phi*pi/180, -1);
% if (inside)
% if ( (x2<=x0)&&(x0<=x2_top)&&(y2<=y0)&&(y0<=y2_top) )
% imseg(x,y) = 1; % good matches / target region
% else
% imseg(x,y) = 0.5;
% end
% end
%
% end
% end
% figure;
% imshow(imseg);
end