-
Notifications
You must be signed in to change notification settings - Fork 3
/
selfsim_cpp.txt
234 lines (234 loc) · 8.58 KB
/
selfsim_cpp.txt
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
--------------------------------------------------------
This code implements the Self-Similarity method of
Shechtman&Irani from CVPR'07, and is for
non-commercial use only (patent pending). For commerical
licensing please refer to YEDA (www.yedarnd.com).
--------------------------------------------------------
This program itself was written by Ori Brostovski,
with algorithmic help from Meirav Galun and Eli Shechtman.
This version was compiled in date Jul 16 2009
and in time 10:26:04.
For more information, see ``README''
For command-line parameters, add --help to the command line
--------------------------------------------------------
Usage 1.1: selfsim_cpp QueryImage QueryMask DataImage DataMask
Usage 1.2: selfsim_cpp QueryImage QueryMask DataImage
Usage 1.3: selfsim_cpp QueryImage DataImage
Usage 2.1: selfsim_cpp --OnlyDesc Image Mask
Usage 2.2: selfsim_cpp --OnlyDesc Image
Usage 3: selfsim_cpp --OnlyVote QueryDesc DataDesc
--------------------------------------------------------
Parameter info (note, some parameters
may contain a list of values, to specify
such a list, simply assign the parameter twice
i.e. --GBSigmas=1 --GBSigmas=1.5):
Name: --Angles
Type: number
Default value: 15
Description: Number of sectors per ring of log-polar bins
----------------------------------------------
Name: --AutoCorrCoef
Type: number
Default value: 4
Description: Used in the normalization of surfaces to say how much differences are significant
----------------------------------------------
Name: --BiasFeatures
Type: flag
Default value: false
Description: Used in the normalization of surfaces to say whether the normalization should be biased
----------------------------------------------
Name: --BinAverageMethod
Type: string
Default value: `max',
Description: Method for binning correlation surfaces (possible values: `max')
----------------------------------------------
Name: --DescOut
Type: string
Default value: `',
Description: Changes the name for the outputted .desc file (use with --OnlyDesc)
----------------------------------------------
Name: --EvaluationCircle
Type: circle
Default value:
Description: For performance evaluation purposes, specifies a circle where we should find a query
----------------------------------------------
Name: --FixedMultiscalePointGenerator
Type: string
Default value: `',
Description: Instead of generating sampling points, gets them from a file
----------------------------------------------
Name: --FixedPointGenerator
Type: string
Default value: `',
Description: Instead of generating interest point, take interest points for the largest scale and re-calculate them for the lower scales note that this means the same number of points in each scale
----------------------------------------------
Name: --GBSigmas
Type: number
Default value: 0, 0.75, 1.5
Description: Sigmas using for geometric blur
----------------------------------------------
Name: --Help
Type: flag
Default value: false
Description: Prints our parameter information
----------------------------------------------
Name: --LocSigma
Type: number
Default value: 5
Description: Sigma of the locational gaussian used for voting also affects it's size
----------------------------------------------
Name: --MaxPoints
Type: number
Default value: 20000
Description: Maximal allowed number of interest points
----------------------------------------------
Name: --MinSim
Type: number
Default value: 0.7
Description: Similarity threshold for pruning features (set below zero to disable)
----------------------------------------------
Name: --NoColorSpaceLAB
Type: flag
Default value: false
Description: Uses RGB instead of LAB
----------------------------------------------
Name: --NoPart
Type: flag
Default value: false
Description: Do not calculate features for surfaces with unknown parts
----------------------------------------------
Name: --NoVoteFinal
Type: flag
Default value: false
Description: Do not output vote_final.png
----------------------------------------------
Name: --OnlyDesc
Type: flag
Default value: false
Description: Only collect descriptors, avoid voting
----------------------------------------------
Name: --OnlyVote
Type: flag
Default value: false
Description: Only vote, avoid collecting descriptors
----------------------------------------------
Name: --OutputContributions
Type: flag
Default value: false
Description: Give some info on how much contribution each query feature gave
----------------------------------------------
Name: --PartMedian
Type: flag
Default value: false
Description: Replace unknown surface areas with median of surrounding data
----------------------------------------------
Name: --PixelNoiseThresh
Type: number
Default value: 2500
Description: Used in the normalization of surfaces as a minimal divisor toget rid of differences due to noise
----------------------------------------------
Name: --PointsPerPixel
Type: number
Default value: 0.03
Description: Determines density of interest point grid
----------------------------------------------
Name: --PreBinningSigma
Type: number
Default value: 0.5
Description: Used for surface blur before log-polar binning
----------------------------------------------
Name: --RPatch
Type: number
Default value: 2
Description: Radius of the patch for calculating descriptors
----------------------------------------------
Name: --RSearch
Type: number
Default value: 38
Description: Radius of the search area around the patch
----------------------------------------------
Name: --RadiusFirst
Type: number
Default value: 6
Description: Radius of the first ring of log-polar bins
----------------------------------------------
Name: --RadiusPower
Type: number
Default value: 0.24
Description: Increasing power applied to RadiusStep
----------------------------------------------
Name: --RadiusStep
Type: number
Default value: 6
Description: Base difference between different radii
----------------------------------------------
Name: --ScaleBase
Type: number
Default value: 1.4142
Description: Multiplicative Difference between scales (for feature collection)
----------------------------------------------
Name: --ScaleFirstPower
Type: number
Default value: 0
Description: First power of ScaleBase coefficient (for feature collection)
----------------------------------------------
Name: --ScaleLastPower
Type: number
Default value: 2
Description: Last power of ScaleBase coefficient (for feature collection)
----------------------------------------------
Name: --ScoreCircle
Type: circle
Default value:
Description: Create a vote score file with feature scores, the center of the given circle is assumed to be the correct peak (using (-1,-1,-1) will assume that the current peak is also the correct one.)
----------------------------------------------
Name: --SimDistThresh
Type: number
Default value: 9
Description: Affects the logistic function used as similarity measure for voting
----------------------------------------------
Name: --SimDistWidth
Type: number
Default value: 7
Description: Affects the logistic function used as similarity measure for voting
----------------------------------------------
Name: --SparseThresh
Type: number
Default value: 0.03
Description: Sparseness threshold for pruning features (set below zero to disable, note that all-equal feature will still be pruned)
----------------------------------------------
Name: --Verbose
Type: flag
Default value: false
Description: Print out more information
----------------------------------------------
Name: --VoteFilterAdd
Type: string
Default value: `',
Description: Allows specifying a file that tells which features should be used (There are two kinds of weights, additive, and multiplicative - note that these are add/mul in log-space.)
----------------------------------------------
Name: --VoteFilterMult
Type: string
Default value: `',
Description: Allows specifying a file that tells which features should be used (There are two kinds of weights, additive, and multiplicative - note that these are add/mul in log-space.)
----------------------------------------------
Name: --VotingScaleBase
Type: number
Default value: -1
Description: Multiplicative Difference between scales (For voting, a negative value implies to vote for all scales.)
----------------------------------------------
Name: --VotingScaleFirstPower
Type: number
Default value: 0
Description: First power of ScaleBase coefficient (For voting.)
----------------------------------------------
Name: --VotingScaleLastPower
Type: number
Default value: 2
Description: Last power of ScaleBase coefficient (For voting.)
----------------------------------------------
Name: --help
Type: flag
Default value: false
Description: Prints our parameter information
----------------------------------------------