forked from BRAINSia/BRAINSStandAlone
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathBRAINSCutCMD.py
executable file
·260 lines (222 loc) · 15.2 KB
/
BRAINSCutCMD.py
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
#!/usr/bin/env python
#
# TODO
# :: copy model file into md5 repository
# :: connect input/output in the BAW
import argparse
import subprocess
def addProbabilityMapElement(probabilityMap, maskName, outputStream):
outputStream.write(" <ProbabilityMap StructureID = \"" + maskName + "\"\n")
outputStream.write(" Gaussian = \"1.0\"\n")
outputStream.write(" GenerateVector = \"true\"\n")
outputStream.write(" Filename = \"" + probabilityMap + "\"\n")
outputStream.write(" />\n")
def xmlGenerator(args, roi=""):
xmlFilename = args.xmlFilename + roi + ".xml"
outputStream = open(xmlFilename, 'w')
registrationID = "BSpline_ROI"
outputStream.write("<AutoSegProcessDescription>\n")
#
# template
#
outputStream.write(" <DataSet Name=\"template\" Type=\"Atlas\" >\n")
outputStream.write(" <Image Type=\"T1\" Filename=\"{fn}\" />\n".format(fn=args.inputTemplateT1))
if args.inputSubjectT2Filename is not None:
outputStream.write(" <Image Type=\"T2\" Filename=\"{fn}\" />\n".format(fn="na"))
outputStream.write(" <Image Type=\"GadSG\" Filename=\"{fn}\" />\n".format(fn="na"))
# outputStream.write( " <Image Type=\"TotalGM\" Filename=\"{fn}\" />\n".format(fn="na"))
# outputStream.write( " <Mask Type=\"RegistrationROI\" Filename=\"{fn}\" />\n".format(fn=args.inputTemplateRegistrationROIFilename))
outputStream.write(" <SpatialLocation Type=\"rho\" Filename=\"" + args.inputTemplateRhoFilename + "\" />\n")
outputStream.write(" <SpatialLocation Type=\"phi\" Filename=\"" + args.inputTemplatePhiFilename + "\" />\n")
outputStream.write(" <SpatialLocation Type=\"theta\" Filename=\"" + args.inputTemplateThetaFilename + "\" />\n")
outputStream.write(" </DataSet>\n")
#
# Registration
#
outputStream.write(" <RegistrationConfiguration \n")
outputStream.write(" ImageTypeToUse = \"T1\"\n")
outputStream.write(" ID = \"" + registrationID + "\"\n")
outputStream.write(" BRAINSROIAutoDilateSize= \"1\"\n")
outputStream.write(" />\n")
#
# training vector configuration (feature vector)
#
outputStream.write(" <NeuralNetParams MaskSmoothingValue = \"0.0\"\n")
outputStream.write(" GradientProfileSize = \"1\"\n")
outputStream.write(" TrainingVectorFilename = \"" + args.trainingVectorFilename + "\"\n")
# outputStream.write( " TrainingModelFilename = \""+args.modelFileBasename+"\"\n")
outputStream.write(" TrainingModelFilename = \"na\"\n")
# outputStream.write( " TrainingModelFilename = \"/nfsscratch/PREDICT/TEST_BRAINSCut/20120828ANNModel_Model_RF100.txt\"\n")
outputStream.write(" TestVectorFilename = \"na\"\n")
if roi == "caudate":
outputStream.write(" Normalization = \"" + 'Linear' + "\"\n")
else:
outputStream.write(" Normalization = \"" + args.vectorNormalization + "\"\n")
outputStream.write(" />\n")
#
# random forest parameters
#
outputStream.write(" <RandomForestParameters \n")
outputStream.write(" MaxDepth= \"1\"\n") # dummy
outputStream.write(" MaxTreeCount= \"1\"\n") # dummy
outputStream.write(" MinSampleCount= \"5\"\n")
outputStream.write(" UseSurrogates= \"false\"\n")
outputStream.write(" CalcVarImportance= \"false\"\n")
outputStream.write(" />\n")
#
# ANN Parameters
#
outputStream.write(" <ANNParameters Iterations = \"5\"\n")
outputStream.write(" MaximumVectorsPerEpoch = \"70000\"\n")
outputStream.write(" EpochIterations = \"100\"\n")
outputStream.write(" ErrorInterval = \"1\"\n")
outputStream.write(" DesiredError = \"0.000001\"\n")
outputStream.write(" NumberOfHiddenNodes = \"100\"\n")
outputStream.write(" ActivationSlope = \"1.0\"\n")
outputStream.write(" ActivationMinMax = \"1.0\"\n")
outputStream.write(" />\n")
#
# apply conditions
#
outputStream.write("<ApplyModel CutOutThresh = \"0.05\"\n")
outputStream.write(" MaskThresh = \"0.5\"\n")
outputStream.write(" GaussianSmoothingSigma = \"0.0\"\n")
outputStream.write(" />\n")
#
# add probability maps (ROIs)
#
if roi == "caudate":
addProbabilityMapElement(args.probabilityMapsLeftCaudate, "l_caudate", outputStream)
addProbabilityMapElement(args.probabilityMapsRightCaudate, "r_caudate", outputStream)
elif roi == 'putamen':
addProbabilityMapElement(args.probabilityMapsLeftPutamen, "l_putamen", outputStream)
addProbabilityMapElement(args.probabilityMapsRightPutamen, "r_putamen", outputStream)
elif roi == 'thalamus':
addProbabilityMapElement(args.probabilityMapsLeftThalamus, "l_thalamus", outputStream)
addProbabilityMapElement(args.probabilityMapsRightThalamus, "r_thalamus", outputStream)
elif roi == 'hippocampus':
addProbabilityMapElement(args.probabilityMapsLeftHippocampus, "l_hippocampus", outputStream)
addProbabilityMapElement(args.probabilityMapsRightHippocampus, "r_hippocampus", outputStream)
elif roi == 'accumben':
addProbabilityMapElement(args.probabilityMapsLeftAccumben, "l_accumben", outputStream)
addProbabilityMapElement(args.probabilityMapsRightAccumben, "r_accumben", outputStream)
elif roi == 'globus':
addProbabilityMapElement(args.probabilityMapsLeftGlobus, "l_globus", outputStream)
addProbabilityMapElement(args.probabilityMapsRightGlobus, "r_globus", outputStream)
#
# subject
#
outputStream.write(" <DataSet Name=\"subject\" Type=\"Apply\"")
outputStream.write(" OutputDir=\"./\" >\n")
outputStream.write(" <Image Type=\"T1\" Filename=\"" + args.inputSubjectT1Filename + "\" />\n")
if args.inputSubjectT2Filename is not None:
outputStream.write(" <Image Type=\"T2\" Filename=\"" + args.inputSubjectT2Filename + "\" />\n")
outputStream.write(" <Image Type=\"GadSG\" Filename=\"" + args.inputSubjectGadSGFilename + "\" />\n")
if roi == "caudate":
outputStream.write(" <Image Type=\"candiateRegion\" Filename=\"" + args.candidateRegion + "\" />\n")
# outputStream.write( " <Image Type=\"TotalGM\" Filename=\"{fn}\" />\n".format(fn=args.inputSubjectTotalGMFilename))
# outputStream.write( " <Mask Type=\"RegistrationROI\" Filename=\"{fn}\" />\n".format(fn=args.inputSubjectRegistrationROIFilename))
# outputStream.write( " <Mask Type=\"l_caudate\" Filename=\""+args.outputBinaryLeftCaudate+"\" />\n")
# outputStream.write( " <Mask Type=\"r_caudate\" Filename=\""+args.outputBinaryRightCaudate+"\" />\n")
# outputStream.write( " <Mask Type=\"l_putamen\" Filename=\""+args.outputBinaryLeftPutamen+"\" />\n")
# outputStream.write( " <Mask Type=\"r_putamen\" Filename=\""+args.outputBinaryRightPutamen+"\" />\n")
# outputStream.write( " <Mask Type=\"l_thalamus\" Filename=\""+args.outputBinaryLeftThalamus+"\" />\n")
# outputStream.write( " <Mask Type=\"r_thalamus\" Filename=\""+args.outputBinaryRightThalamus+"\" />\n")
# outputStream.write( " <Mask Type=\"l_hippocampus\" Filename=\""+args.outputBinaryLeftHippocampus+"\" />\n")
# outputStream.write( " <Mask Type=\"r_hippocampus\" Filename=\""+args.outputBinaryRightHippocampus+"\" />\n")
# outputStream.write( " <Mask Type=\"l_accumben\" Filename=\""+args.outputBinaryLeftAccumben+"\" />\n")
# outputStream.write( " <Mask Type=\"r_accumben\" Filename=\""+args.outputBinaryRightAccumben+"\" />\n")
# outputStream.write( " <Mask Type=\"l_globus\" Filename=\""+args.outputBinaryLeftGlobus+"\" />\n")
# outputStream.write( " <Mask Type=\"r_globus\" Filename=\""+args.outputBinaryRightGlobus+"\" />\n")
# if args.inputSubjectBrainMaskFilename != "NA":
# outputStream.write( " <Mask Type=\"RegistrationROIi\" Filename=\""+args.inputSubjectBrainMaskFilename+"\" />\n")
if not args.deformationFromSubjectToTemplate is None:
outputStream.write(' <Registration SubjToAtlasRegistrationFilename="' + args.deformationFromSubjectToTemplate + '"\n')
else:
outputStream.write(' <Registration SubjToAtlasRegistrationFilename="" \n')
outputStream.write(" AtlasToSubjRegistrationFilename=\"" + args.deformationFromTemplateToSubject + "\"\n")
outputStream.write(" ID=\"" + registrationID + "\" /> \n")
outputStream.write(" </DataSet>\n")
outputStream.write("</AutoSegProcessDescription>\n")
outputStream.close()
return xmlFilename
##
## main
##
brainscutParser = argparse.ArgumentParser(description='BRAINSCut command line argument parser')
# HACK: This is to allow special treatment of caudates with masking
brainscutParser.add_argument('--candidateRegion', help='Specify the valid candidate region for caudate', required=True)
#
# input arguments
#
brainscutParser.add_argument('--inputSubjectT1Filename', help='T1 subject filename', required=True)
brainscutParser.add_argument('--inputSubjectT2Filename', help='T2 subject filename', required=False)
# brainscutParser.add_argument('--inputSubjectTotalGMFilename', help='TotalGM filename', required=True )
brainscutParser.add_argument('--inputSubjectGadSGFilename', help='GadSG subject filename', required=False)
# brainscutParser.add_argument('--inputSubjectBrainMaskFilename', help='BrainMask subject filename' )
# brainscutParser.add_argument('--inputSubjectRegistrationROIFilename', help='template brain mask filename' )
brainscutParser.add_argument('--inputTemplateT1', help='template T1-weighted filename', required=True)
# brainscutParser.add_argument('--inputTemplateRegistrationROIFilename', help='template brain region filename', required=True )
brainscutParser.add_argument('--inputTemplateRhoFilename', help='template rho filename', required=True)
brainscutParser.add_argument('--inputTemplatePhiFilename', help='template phi filename', required=True)
brainscutParser.add_argument('--inputTemplateThetaFilename', help='template theta filename', required=True)
brainscutParser.add_argument('--trainingVectorFilename', help='training vector filename', default="NA")
# brainscutParser.add_argument('--modelFileBasename', help='model filei base name for net configuration file (xml).', default="NA" )
brainscutParser.add_argument('--modelFilename', help='model filename', default="NA", required=True)
brainscutParser.add_argument('--vectorNormalization', help='feature vector normalization (IQR,Linear,Sigmoid_Q01,Sigmoid_Q05,ZScore,NONE)', required=True)
# probability maps
brainscutParser.add_argument('--probabilityMapsLeftCaudate', help='model probability maps for left caudate', required=True)
brainscutParser.add_argument('--probabilityMapsRightCaudate', help='model probability maps for right caudate', required=True)
brainscutParser.add_argument('--probabilityMapsLeftPutamen', help='model probability maps for left putamen', required=True)
brainscutParser.add_argument('--probabilityMapsRightPutamen', help='model probability maps for right putamen', required=True)
brainscutParser.add_argument('--probabilityMapsLeftThalamus', help='model probability maps for left thalamus', required=True)
brainscutParser.add_argument('--probabilityMapsRightThalamus', help='model probability maps for right thalamus', required=True)
brainscutParser.add_argument('--probabilityMapsLeftHippocampus', help='model probability maps for left hippocampus', required=True)
brainscutParser.add_argument('--probabilityMapsRightHippocampus', help='model probability maps for right hippocampus', required=True)
brainscutParser.add_argument('--probabilityMapsLeftAccumben', help='model probability maps for left accumben', required=True)
brainscutParser.add_argument('--probabilityMapsRightAccumben', help='model probability maps for right accumben', required=True)
brainscutParser.add_argument('--probabilityMapsLeftGlobus', help='model probability maps for left globus', required=True)
brainscutParser.add_argument('--probabilityMapsRightGlobus', help='model probability maps for right globus', required=True)
brainscutParser.add_argument('--deformationFromTemplateToSubject', help="deformationFromTemplateToSubject")
brainscutParser.add_argument('--deformationFromSubjectToTemplate', help="deformationFromSubjectToTemplate")
#
# output arguments
#
brainscutParser.add_argument('--outputBinaryLeftCaudate', help='output binary file name for left caudate')
brainscutParser.add_argument('--outputBinaryRightCaudate', help='output binary file name for right caudate')
brainscutParser.add_argument('--outputBinaryLeftPutamen', help='output binary file name for left putamen')
brainscutParser.add_argument('--outputBinaryRightPutamen', help='output binary file name for right putamen')
brainscutParser.add_argument('--outputBinaryLeftThalamus', help='output binary file name for left thalamus')
brainscutParser.add_argument('--outputBinaryRightThalamus', help='output binary file name for right thalamus')
brainscutParser.add_argument('--outputBinaryLeftHippocampus', help='output binary file name for left hippocampus')
brainscutParser.add_argument('--outputBinaryRightHippocampus', help='output binary file name for right hippocampus')
brainscutParser.add_argument('--outputBinaryLeftAccumben', help='output binary file name for left accumben')
brainscutParser.add_argument('--outputBinaryRightAccumben', help='output binary file name for right accumben')
brainscutParser.add_argument('--outputBinaryLeftGlobus', help='output binary file name for left globus')
brainscutParser.add_argument('--outputBinaryRightGlobus', help='output binary file name for right globus')
brainscutParser.add_argument('--xmlFilename', help='BRAINSCut xml configuration filename', default="output.xml")
args = brainscutParser.parse_args()
## HACK: DOUBLE CHECK THAT IQR IS USED
if args.vectorNormalization != "IQR":
print "ERROR: ONLY IQR SUPPORTED AT THE MOMENT"
exit - 1
print(args)
roiList = ['accumben', 'caudate', 'putamen', 'globus', 'thalamus', 'hippocampus']
for roi in roiList:
currentXmlFilename = xmlGenerator(args, roi)
if roi == "caudate":
currentModelFilename = args.modelFilename[:-3] + '_' + roi + '_LinearWithMask.gz' # trainModelFile.txtD0060NT0060_caudate_LinearWithMask.gz
else:
currentModelFilename = args.modelFilename[:-3] + '_' + roi + '.gz' # trainModelFile.txtD0060NT0060_accumben.gz
BRAINSCutCommand = ["BRAINSCut" + " --applyModel " +
" --netConfiguration " + currentXmlFilename +
" --modelFilename " + currentModelFilename +
" --method RandomForest" +
" --numberOfTrees 60 --randomTreeDepth 60"
]
print("HACK: BRAINCUT COMMAND: {0}".format(BRAINSCutCommand))
subprocess.call(BRAINSCutCommand, shell=True)
"""
script to be run
BRAINSCut --applyModel --netConfiguration BRAINSTools-build/BRAINSCut/TestSuite/TestSuite/NetConfigurations/output.xml --modelFilename TrainedModels/20110919ANNModel_allSubcorticals.txtD0050NT0050 --method RandomForest
"""