Re: [help-GIFT] Re: Features Extraction

[Epa Uwimana]

I read papers and I understand what kinds of features that GIFT 
extracts but the problem I am having is about the unpacking files that 
are created when I extract features. Here is a good example: I have 
a .fts file and when I read it I get something like this 
63:108
63:123
63:132
63:178
63:188
63:197
220:14
15:1
15:1
15:1
15:1
16:1
16:1
16:1
17:1
17:1
17:1
1:0
1:0
1:0
128:63
128:63
128:63
128:63
128:63
128:63
63:89
63:99
63:109
63:220
230:22
0:0
113:23
123:23
224:23
0:0
63:89
63:97
63:107
63:35
63:46
63:56
63:67
63:77
63:87
63:97
63:106
63:117
63:127
63:137
63:147
63:157
63:168
63:177
63:187
63:55
63:65
63:75
63:85
63:145
63:156
63:49
63:59
63:159
63:167
63:1
63:145
63:205
215:31
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:0
128:63
1:0
1:0
1:0
1:0
1:0
1:0
1:0
0:0
128:63
34:1
34:1
34:1
34:1
34:1
34:1
34:1
34:1
34:1
34:1
1:0
0:0
0:0
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
64:38
143:38
81:40
242:40
5:41
135:41
165:41
175:41
71:42
79:42
92:42
109:42
119:42
129:42
230:42
0:128
43:1
43:1
43:1
43:1
43:1
0:0
0:128
57:44
77:44
87:44
97:44
107:44
117:44
177:44
37:45
81:45
189:45
199:45
46:1
46:1
46:1
46:1
46:1
0:0
6:47
89:47
120:47
129:47
156:47
167:47
179:47
249:47
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:128
63:49
63:57
63:87
63:123
63:133
63:137
63:147
63:197
212:49
1:0
0:0
26:50
37:50
106:50
118:50
125:50
141:50
155:50
187:50
197:50
1:0
0:0
63:8
63:19
63:26
63:45
63:57
63:92
63:102
63:107
63:115
63:157
63:168
63:176
63:185
63:207
216:51
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
63:229
0:0
3:53
13:53
25:53
60:53
71:53
84:53
120:53
124:53
133:53
143:53
176:53
188:53
220:53
0:0
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
63:207
222:54
55:1
55:1
55:1
55:1
0:128
0:128
0:128
0:128
0:128
0:128
0:128
0:128
128:63
63:63
63:87
63:97
63:127
63:198
207:59
0:128
0:128
0:128
0:128
0:128
0:128
0:128
0:128
61:1
61:1
61:1
61:1
61:1
61:1
1:0
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:128
128:63
239:62
4:63
43:63
117:63
125:63
133:63
183:63
78:64
90:64
127:64
237:64
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
63:63
63:93
63:47
63:251
0:0
0:0
0:0
0:0
0:0
0:128
128:63
49:72
59:72
79:72
90:72
100:72
114:72
127:72
134:72
210:72
1:0
0:0
63:253
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:0
128:63
243:73
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:0
0:128
128:63
235:74
0:128
75:1
75:1
75:1
75:1
76:1
76:1
77:1
77:1
77:1
1:0
0:0
0:0
0:0
0:0
0:0
0:0
63:254
0:128
0:128
0:128
0:128
0:128
128:63
128:63
128:63
128:63
128:63
128:63
128:63
128:63
245:82
63:6
63:14
63:24
63:33
63:147
63:157
63:163
63:175
63:47
63:59
63:80
63:87
63:127
63:179
63:188
63:217
21:85
30:85
85:1
1:0
0:129
0:193
201:200
128:59
60:40
1:0
1:0
0:161
0:129
0:193
193:192
176:48
136:136
128:0
61:85
70:85
71:85
0:137
0:225
192:60
1:0
128:0
93:63
176:61
16:61
60:90
63:91
61:92
61:93
94:85
85:1
85:1
1:0
0:129
0:217
209:208
48:61
128:59
128:59
12:63
28:62
61:113
85:1
85:1
85:1
0:177
0:225
160:32
68:63
20:62
0:61
124:85
125:85
126:85
130:85
131:85
1:0
1:0
1:0
1:0
1:0
0:225
128:128
200:61
141:85
142:85
143:85
144:85
145:85
147:85

I can see that the first three digits are for Feature Id and the 
second part is frequency (Thanks for explanation). Now how do I go 
from here to know what feature ID is (144 or 145). I tried to read the 
code about Features extraction but I am not sure why I am missing how 
I can connect this kind of output to Inverted Files. Please let me 
know if I am missing something. Sometimes is not easy to fully 
understand another person's code. My automate goal to be able to know 
how many features I extracted from the image and what they are.
I hope I am not being a pain but I am sure this info is somewhere in 
this code but I am missing it. Even just pseudocode on how this is 
done might help. Just out of curiosity, does it matter that my 
collection is not colored?

Thanks again

Epa Uwimana


Quoting Henning Müller <address@hidden>:

> Hi,
> 
> the papers expalin pretty well what kind of features exist. Then you
> can 
> create a list of feature number vs exact feature description as
> Wolfgang 
> explained in his mail . The executable to do so is write_feature_desc
> or 
> something like that in the FeatureExtraction directory and it
> requires 
> as input a ppm image of the size 256x256. Have you tried this?
> 
> Cheers, Henning
> 
> Epa wrote:
> > I read some papers about GIFT back in 2000 and 1999 and I still
> can't find 
> > what values stand for HSV, GABOR, COLORMAP or SIZE from the out of
> this script
> > 
> > 
> > 
> > 
> > 
> > 
> > _______________________________________________
> > help-GIFT mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/help-gift
> > 
> 
> -- 
>   ---------------------------------------
>   Henning Müller (PhD)
>   University & Hospitals of Geneva
>   Service of Medical Informatics
>   24, rue Micheli-du-Crest
>   CH-1211 Geneva 14, Switzerland
>   Tel  +41 22 372-6175
>   Fax  +41 22 372-8680
>   email address@hidden
> 
>