forked from torrvision/spaint
-
Notifications
You must be signed in to change notification settings - Fork 0
/
README
212 lines (145 loc) · 6.47 KB
/
README
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
This is the software bundle "SemanticPaint", created by:
Stuart Golodetz <[email protected]>
Michael Sapienza <[email protected]>
Tommaso Cavallari <[email protected]>
Nicholas A Lord <[email protected]>
Julien Valentin <[email protected]>
Vibhav Vineet <[email protected]>
Ming-Ming Cheng <[email protected]>
Anurag Arnab <[email protected]>
Stephen Hicks <[email protected]>
Shahram Izadi <[email protected]>
Philip H S Torr <[email protected]>
It is built on top of InfiniTAM v3.5, a highly-efficient,
open-source 3D reconstruction engine developed by Oxford's
Active Vision Group. For more details about InfiniTAM v3.5,
please get in touch with one of the following people:
Victor Adrian Prisacariu <[email protected]>
Olaf Kaehler <[email protected]>
Stuart Golodetz <[email protected]>
Tommaso Cavallari <[email protected]>
1. Building the System
----------------------
1.1 Overview
SemanticPaint builds with CMake 3.8.2 and above on Windows, Ubuntu and Mac OS X.
It has been tested on recent versions of all three operating systems, but may
require additional work to build on older versions.
1.2 Dependencies
SemanticPaint depends on various 3rd party libraries. Some of these are
required, in the sense that SemanticPaint will not function without them;
others are optional, but skipping them will reduce functionality.
Suitable versions of the required dependencies and some of the optional
dependencies have been included in the repository, so SemanticPaint should
build "out of the box". However, most users will probably want to install
the remaining optional libraries in order to enable full functionality.
- ArrayFire (version 3.3.2)
Status: Optional (needed for touch interaction and median filtering)
Default: Disabled
Flag: WITH_ARRAYFIRE
- Boost (version 1.58.0)
Status: Required
- CUDA (version 7.5 or above)
Status: Optional (needed for GPU acceleration)
Default: Enabled
Flag: WITH_CUDA
- Eigen (version 3.2.2)
Status: Required
- GLEW (version 1.12.0)
Status: Required on Windows/Ubuntu
- InfiniTAM (version 3.5)
Status: Required
- Leap Motion SDK (version 2.2.1.24116)
Status: Optional (needed for experimental Leap Motion support)
Default: Disabled
Flag: WITH_LEAP
- Oculus SDK (version 0.5.0.1)
Status: Optional (needed for Oculus Rift support)
Default: Disabled
Flag: WITH_OVR
- OpenCV (version 3.1.0)
Status: Optional (needed for feature inspection mode)
Default: Disabled
Flag: WITH_OPENCV
- OpenGL
Status: Required
- OpenMP
Status: Optional, but recommended (needed for faster training/prediction)
Default: Disabled
Flag: WITH_OPENMP
Notes: Doesn't work on Mac OS X
- OpenNI (version 2)
Status: Optional, but recommended (needed for live reconstruction)
Default: Disabled
Flag: WITH_OPENNI
- SDL (version 2-2.0.7)
Status: Required
- Vicon SDK
Status: Optional (needed for the Vicon tracker)
Default: Disabled
Flag: WITH_VICON
1.3 Build Process
Common steps on all platforms:
a) Choose a root directory, hereafter referred to as <root>.
b) Clone the InfiniTAM repository into <root>/InfiniTAM.
c) Build the infinitam_v3.5 branch of InfiniTAM using CMake (using <root>/InfiniTAM/InfiniTAM/build as your build directory).
d) Clone SemanticPaint into <root>/spaint, e.g.
$ git clone [email protected]:torrvision/spaint.git spaint
e) Change to the spaint directory:
$ cd spaint
1.3.1 Build Process (Windows)
f) Using a Bash shell (e.g. Git Bash, provided with msysgit), run the top-level build script, e.g. for VS2013:
$ ./build-win.sh 12 Release
g) Run the CMake GUI and specify <root>/spaint/build as the build directory.
h) Enable any optional libraries you want and supply the paths.
i) Reconfigure and regenerate the Visual Studio solution.
j) Build with the optional libraries enabled (just run the above build command again).
1.3.2 Build Process (Unix)
f) Run the top-level build script:
$ ./build-nix.sh "Unix Makefiles" Release
g) Change to the build directory and run ccmake:
$ cd build
$ ccmake ..
h) Enable any optional libraries you want and supply the paths.
i) Reconfigure and regenerate the makefile.
j) Build with the optional libraries enabled, e.g.
$ make -j4
2. Running the System
---------------------
The build process will build various applications, libraries and test
programs in <root>/build. The main application can be found at:
<root>/build/bin/apps/spaintgui/spaintgui
It can be run "out of the box" from any working directory for live
reconstruction, provided you have built with OpenNI support.
If you want to reconstruct from data stored on disk (e.g. the Teddy
sequence that is used in InfiniTAM), you can call the application
as follows:
$ ./spaintgui -c <path>/Teddy/calib.txt -r <path>/Teddy/Frames/%04i.ppm -d <path>/Teddy/Frames/%04i.pgm
The arguments specify a text file containing calibration parameters,
and masks for the RGB and depth images in the input sequence.
Note that spaintgui also supports its own saving/loading of sequences
to/from the <root>/build/bin/apps/spaintgui/sequences subdirectory.
These can be loaded straightforwardly via:
$ ./spaintgui -s <sequence name>
3. Additional Documentation
---------------------------
Further details about the original concept for SemanticPaint can be
found in the following paper:
SemanticPaint: Interactive 3D Labeling and Learning at your Fingertips.
ACM Transactions on Graphics, 34(5), August 2015.
There is also a technical report that describes this implementation of
the SemanticPaint concept:
SemanticPaint: A Framework for the Interactive Segmentation of 3D Scenes.
arXiv e-print 1510.03727, October 2015.
4. Troubleshooting Tips
-----------------------
If you have any trouble with the build, here are some of the likely causes:
* The CMake executable is not on your path.
* The MSBuild executable is not on your path. (Windows)
* The Visual C++ compiler (cl.exe) is not on your path. (Windows)
* The latest Visual Studio updates have not been installed. (Windows)
* You're using a version of CMake that has a CUDA compilation bug (e.g. an early version of 3.11). Try CMake 3.9 instead. (Windows)
* You built Boost other than with the build script (which contains some bug fixes). Please use the build script :)
History:
2017-AUG-08: second public release
2015-SEP-26: first public release
2015-JUL-23: initial draft