-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathfod_cnn_genData.py
434 lines (330 loc) · 20.3 KB
/
fod_cnn_genData.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
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sat Aug 20 08:25:14 2022
@author: ch209389
"""
import os
from os import listdir
from os.path import isdir, join
import numpy as np
from shutil import copyfile
import nibabel as nib
import pandas as pd
import time
import dipy.core.sphere as dipysphere
import SimpleITK as sitk
import crl_aux
import dk_seg
###############################################################################
# Generate the training data
dhcp_dir = '/home/ch051094/Desktop/dhcp_4_hamza/'
subj_info = pd.read_csv(dhcp_dir + 'participants.tsv', delimiter='\t')
# to save the images that will be used for model training
base_dir = '/home/ch051094/Desktop/DataDWI2FOD/'
images_dir = base_dir + 'images/'
training_dir = base_dir + 'DL_90sub_EqualSampling9-TestSet15-RandomBvecsCondNumber/'
mapInputToSH = True
os.makedirs(training_dir, exist_ok=True)
prefix = '' # '/home/ch051094/anaconda3/bin/' #to be added for right mrtrix path
n_max_subjects = 100 # 73
subj_count = 0
min_age = 25
max_age = 50 # 38
desired_spacing = (1.0, 1.0, 1.0)
# number of measurement subsets
# N_grad_1 = 32 #b400
# Xp, Yp, Zp= crl_aux.distribute_on_hemisphere_spiral(N_grad_1)
# sphere_grad = dipysphere.Sphere(Xp, Yp, Zp)
# v_grad_1, _ = sphere_grad.vertices, sphere_grad.faces
#
# N_grad_2 = 44 #b1000
# Xp, Yp, Zp= crl_aux.distribute_on_hemisphere_spiral(N_grad_2)
# sphere_grad = dipysphere.Sphere(Xp, Yp, Zp)
# v_grad_2, _ = sphere_grad.vertices, sphere_grad.faces
#
# N_grad_3 = 64 #b2600
# Xp, Yp, Zp= crl_aux.distribute_on_hemisphere_spiral(N_grad_3)
# sphere_grad = dipysphere.Sphere(Xp, Yp, Zp)
# v_grad_3, _ = sphere_grad.vertices, sphere_grad.faces
#
# N_grad_list= [ N_grad_1, N_grad_2, N_grad_3 ]
# V_grad_list= [ v_grad_1, v_grad_2, v_grad_3 ]
N_grad_1 = 6
Xp, Yp, Zp = crl_aux.optimized_six()
sphere_grad = dipysphere.Sphere(Xp, Yp, Zp)
v_grad_1, _ = sphere_grad.vertices, sphere_grad.faces
N_grad_list = [N_grad_1]
V_grad_list = [v_grad_1]
temp = []
start_time = time.time()
for subj_ind in range(n_max_subjects):
subj = subj_info['participant_id'][subj_ind]
print('Processing subject ', subj)
#anat_dir= dhcp_dir + 'dhcp_anat_pipeline/sub-' + subj
dmri_dir = dhcp_dir + 'dhcp_dmri_pipeline/sub-' + subj
dmri_sess = [d for d in listdir(dmri_dir) if isdir(join(dmri_dir, d))]
sess_tsv = dmri_dir + '/sub-' + subj + '_sessions.tsv' #anat_dir + '/sub-' + subj + '_sessions.tsv'
sess_info= pd.read_csv( sess_tsv , delimiter= '\t')
if len(dmri_sess)==0 or not sess_info.loc[0, 'session_id']== int(dmri_sess[0].split('ses-')[1]):
continue
else:
age_c= sess_info.loc[0, 'scan_age']
if age_c<min_age or age_c>max_age:
continue
else:
if subj_count < n_max_subjects:
subj_count+= 1
print('Reading data for Subject ' + str(subj_ind) + '.')
subject= 'sub-' + subj
session= 'ses-'+ str(sess_info.loc[0, 'session_id'])
dwi_dir = dmri_dir + '/' + session + '/' + 'dwi/'
#ant_dir = anat_dir + '/' + session + '/' + 'anat/'
subject_results_dir= images_dir + subj+'/' #+ '_test/'
os.makedirs(subject_results_dir)
os.chdir(subject_results_dir)
# copy the main diffusion volume
file_name= subject + '_' + session + '_desc-preproc_dwi.nii.gz'
copyfile( dwi_dir + file_name , subject_results_dir + 'dwi.nii.gz' )
# denoising
# my_command= 'dwidenoise dwi.nii.gz dwi_denoised.nii.gz -noise noise.nii.gz -quiet'
# os.system(my_command)
# read the b values/vectors and copy them into the subject directory as well
file_name= subject + '_' + session + '_desc-preproc_dwi'
b_vals= np.loadtxt( dwi_dir + file_name + '.bval' )
b_vecs= np.loadtxt( dwi_dir + file_name + '.bvec' )
copyfile( dwi_dir + file_name + '.bval', subject_results_dir + 'bvals.bval' )
copyfile( dwi_dir + file_name + '.bvec', subject_results_dir + 'bvecs.bvec' )
file_name= subject + '_' + session + '_desc-preproc_space-dwi_brainmask.nii.gz'
copyfile( dwi_dir + file_name , subject_results_dir + 'mask_orig.nii.gz' )
# bias correction
#os.system('FSLDIR=/opt/el7/pkgs/fsl/fsl-5.0.11/bin/fsl:$FSLDIR')
#os.system('export FSLDIR')
my_command= 'dwibiascorrect fsl dwi.nii.gz dwi_debiased.nii.gz -fslgrad bvecs.bvec bvals.bval -quiet'
my_command+= ' -mask mask_orig.nii.gz '
my_command+= ' -bias estimated_bias.nii.gz '
os.system(my_command)
os.remove(subject_results_dir + 'dwi.nii.gz')
# read the denoised volume
dmri_volume= sitk.ReadImage( subject_results_dir + 'dwi_debiased.nii.gz')
dmri_volume= sitk.GetArrayFromImage(dmri_volume)
dmri_volume= np.transpose( dmri_volume, [3,2,1,0] )
# interpolate the data
brain_mask_img= sitk.ReadImage( subject_results_dir + 'mask_orig.nii.gz' )
original_spacing= brain_mask_img.GetSpacing()
original_size= brain_mask_img.GetSize()
n_dwi= dmri_volume.shape[-1]
SX, SY, SZ= [int(original_size[i]/desired_spacing[i]*original_spacing[i]) for i in range(3)]
d_img0= dmri_volume.copy()
dmri_volume= np.zeros( (SX, SY, SZ,n_dwi), np.float32 )
for i in range(n_dwi):
temp= d_img0[:,:,:,i].copy()
temp= np.transpose( temp, [2,1,0] )
temp= sitk.GetImageFromArray(temp)
temp.SetDirection( brain_mask_img.GetDirection() )
temp.SetOrigin( brain_mask_img.GetOrigin() )
temp.SetSpacing( brain_mask_img.GetSpacing() )
temp= dk_seg.resample3d(temp, original_spacing, desired_spacing, sitk.sitkBSpline)
temp= sitk.GetArrayFromImage(temp)
temp= np.transpose( temp, [2,1,0] )
dmri_volume[:,:,:,i]= temp
del d_img0
# Read and resample the brain mask
brain_mask_img= dk_seg.resample3d(brain_mask_img, original_spacing, desired_spacing, sitk.sitkNearestNeighbor)
brain_mask= sitk.GetArrayFromImage(brain_mask_img)
brain_mask= np.transpose( brain_mask, [2,1,0] )
# the directiona, origin, and spacig of the mask, which may be needed below
ref_dir= brain_mask_img.GetDirection()
ref_org= brain_mask_img.GetOrigin()
ref_spc= brain_mask_img.GetSpacing()
# Compute an "eroded" mask to remove some of the boundary voxels
skull_eroded= crl_aux.skull_from_brain_mask(brain_mask, radius= 2.0)
b0_img= dmri_volume[:,:,:,b_vals==0]
b0_img= np.mean(b0_img, axis=-1)
compute_mask= np.logical_and( brain_mask==1, b0_img>1)
compute_mask_eroded= np.logical_and( skull_eroded==0, compute_mask>0)
compute_mask_img= np.transpose(compute_mask, [2,1,0])
compute_mask_img= sitk.GetImageFromArray(compute_mask_img.astype(np.float64))
compute_mask_img.SetDirection(ref_dir)
compute_mask_img.SetOrigin(ref_org)
compute_mask_img.SetSpacing(ref_spc)
sitk.WriteImage(compute_mask_img, subject_results_dir + 'compute_mask_img.nii.gz' )
compute_mask_eroded_img= np.transpose(compute_mask_eroded, [2,1,0])
compute_mask_eroded_img= sitk.GetImageFromArray(compute_mask_eroded_img.astype(np.float64))
compute_mask_eroded_img.SetDirection(ref_dir)
compute_mask_eroded_img.SetOrigin(ref_org)
compute_mask_eroded_img.SetSpacing(ref_spc)
sitk.WriteImage(compute_mask_eroded_img, subject_results_dir + 'compute_mask_eroded_img.nii.gz' )
#
# # seelct subsets of the data for diffusion tensor estimation and fiber orientation estimation
# ind_0_1000= np.where( np.logical_or(b_vals==0, b_vals==1000) )[0]
# b_vals_0_1000= b_vals[ind_0_1000]
# b_vecs_0_1000= b_vecs[:,ind_0_1000]
# gtab_0_1000 = gradient_table(b_vals_0_1000, b_vecs_0_1000)
# d_img_0_1000= dmri_volume[:,:,:,ind_0_1000]
#
# # ind_0_2600= np.where( np.logical_or(b_vals==0, b_vals==2600) )[0]
# # b_vals_0_2600= b_vals[ind_0_2600]
# # b_vecs_0_2600= b_vecs[:,ind_0_2600]
# # gtab_0_2600 = gradient_table(b_vals_0_2600, b_vecs_0_2600)
# # d_img_0_2600= dmri_volume[:,:,:,ind_0_1000]
#
#
temp= nib.load( subject_results_dir + 'compute_mask_img.nii.gz' )
temp.header['srow_x']= temp.affine[0,:]
temp.header['srow_y']= temp.affine[1,:]
temp.header['srow_z']= temp.affine[2,:]
affine= temp.affine
#
# ###################################################################
#
#
# # DTI
#
# np.savetxt(subject_results_dir + 'b_vecs_0_1000.txt', b_vecs_0_1000, delimiter=' ')
# np.savetxt(subject_results_dir + 'b_vals_0_1000.txt', b_vals_0_1000, delimiter=' ')
# os.rename('b_vecs_0_1000.txt', 'b_0_1000.bvec')
# os.rename('b_vals_0_1000.txt', 'b_0_1000.bval')
#
# array_img = nib.Nifti1Image(d_img_0_1000, affine)
# nib.save(array_img, subject_results_dir + 'd_img_0_1000.nii.gz' )
#
# my_command= prefix+'dwi2tensor d_img_0_1000.nii.gz tensor_mrtrix.nii.gz' \
# + ' -mask compute_mask_img.nii.gz -fslgrad b_0_1000.bvec b_0_1000.bval -quiet'
# os.system(my_command)
#
# my_command= prefix+'tensor2metric tensor_mrtrix.nii.gz' \
# + ' -adc MD_mrtrix.nii.gz -fa FA_mrtrix.nii.gz -quiet'
# os.system(my_command)
#
# os.remove(subject_results_dir + 'd_img_0_1000.nii.gz')
# os.remove(subject_results_dir + 'b_0_1000.bvec')
# os.remove(subject_results_dir + 'b_0_1000.bval')
#MSMT-CSD to generate groud-truth FOD
temp = nib.Nifti1Image(dmri_volume, affine)
nib.save(temp, subject_results_dir + 'd_img.nii.gz' )
my_command= prefix+'dwi2response dhollander d_img.nii.gz -fslgrad bvecs.bvec bvals.bval ' + \
' -mask compute_mask_eroded_img.nii.gz response_wm.txt response_gm.txt response_csf.txt -quiet'
os.system(my_command)
my_command= prefix+'dwi2fod msmt_csd d_img.nii.gz -fslgrad bvecs.bvec bvals.bval '
my_command+= ' -mask compute_mask_img.nii.gz '
my_command+= ' response_wm.txt wm.nii.gz '
my_command+= ' response_gm.txt gm.nii.gz '
my_command+= ' response_csf.txt csf.nii.gz -quiet'
os.system(my_command)
os.remove(subject_results_dir + 'd_img.nii.gz')
os.remove(subject_results_dir + 'response_wm.txt')
os.remove(subject_results_dir + 'response_gm.txt')
os.remove(subject_results_dir + 'response_csf.txt')
# Select data subset MULTISHELL
b_vals_all = [400, 1000, 2600]
l=-1
for N_grad, v_grad in zip(N_grad_list, V_grad_list):
l=l+1
#print('N_grad', N_grad)
#print('b_vals_all[l]', b_vals_all[l])
b_vecs_to_chose_from = b_vecs[:, b_vals == b_vals_all[l]]
grad_ind_norot, grad_ang_norot, _ = crl_aux.find_closest_bvecs_no_rotation(v_grad.T, \
b_vecs_to_chose_from.T,
antipodal=True)
#Useful for code of half GS
##print('grad_ind_norot',grad_ind_norot)
#indices_to_chose_from = range(N_grad*2)
##print('b_vecs_to_chose_from', b_vecs_to_chose_from)
#grad_ind_norot = np.setdiff1d(indices_to_chose_from,grad_ind_norot).astype(int)
##print('grad_ind_norot', grad_ind_norot)
if not len(np.unique(grad_ind_norot)) == N_grad:
print('Non-unique directions selected')
print(N_grad-len(np.unique(grad_ind_norot)) )
#else:
print('mean and std of angles %.3f ' % grad_ang_norot.mean(),
' %.3f ' % grad_ang_norot.std(), )
DWI_dhcp_norot = dmri_volume[:, :, :, b_vals == b_vals_all[l]][:, :, :, grad_ind_norot].copy()
for i in range(N_grad):
temp = DWI_dhcp_norot[:, :, :, i] / b0_img
temp[compute_mask < 0.5] = 0
DWI_dhcp_norot[:, :, :, i] = temp.copy()
DWI_dhcp_norot[DWI_dhcp_norot < 0] = 0
DWI_dhcp_norot[DWI_dhcp_norot > 1] = 1
DWI_dhcp_norot = nib.Nifti1Image(DWI_dhcp_norot, affine)
nib.save(DWI_dhcp_norot, subject_results_dir + 'd_mri_subset_' + str(N_grad) + '.nii.gz')
b_vals_selected_curr = b_vals[b_vals == b_vals_all[l]][grad_ind_norot]
b_vecs_selected_curr = b_vecs[:, b_vals == b_vals_all[l]][:, grad_ind_norot]
if l==0:
b_vals_0 = b_vals[b_vals == 0][10:]
#print('b_vecs.shape', b_vecs.shape)
b_vecs_0 = b_vecs[:, b_vals == 0][:,10:]
#print('b_vecs_0.shape',b_vecs_0.shape)
b_vals_selected = np.hstack((b_vals_0, b_vals_selected_curr))
b_vecs_selected = np.hstack((b_vecs_0, b_vecs_selected_curr))
else:
b_vals_selected = np.hstack((b_vals_selected, b_vals_selected_curr))
b_vecs_selected = np.hstack((b_vecs_selected, b_vecs_selected_curr))
if l==0:
d_img_0 = dmri_volume[:, :, :, b_vals == 0][...,10:]
d_img_selected_curr = dmri_volume[:, :, :, b_vals == b_vals_all[l]][:, :, :, grad_ind_norot]
d_img_selected = np.concatenate((d_img_0, d_img_selected_curr), axis=-1)
else:
d_img_selected_curr = dmri_volume[:, :, :, b_vals == b_vals_all[l]][:, :, :, grad_ind_norot]
#print('d_img_selected_curr.shape ',d_img_selected_curr.shape)
d_img_selected = np.concatenate((d_img_selected, d_img_selected_curr), axis=-1)
np.savetxt(subject_results_dir + 'b_vals_selected.txt', b_vals_selected_curr, delimiter=' ')
np.savetxt(subject_results_dir + 'b_vecs_selected.txt', b_vecs_selected_curr, delimiter=' ')
os.rename('b_vals_selected.txt', 'b_vals_selected.bval')
os.rename('b_vecs_selected.txt', 'b_vecs_selected.bvec')
d_img_selected_nib = nib.Nifti1Image(d_img_selected, affine)
nib.save(d_img_selected_nib, subject_results_dir + 'd_img_selected.nii.gz')
my_command = prefix + 'dwi2response dhollander d_img2_selected.nii.gz -fslgrad b_vecs2_selected.bvec b_vals2_selected.bval ' + \
' -mask compute_mask_eroded_img.nii.gz response_wm.txt response_gm.txt response_csf.txt -quiet'
os.system(my_command)
my_command = prefix + 'dwi2fod msmt_csd d_img2_selected.nii.gz -fslgrad b_vecs2_selected.bvec b_vals2_selected.bval '
my_command+= ' -mask compute_mask_img.nii.gz '
my_command+= ' response_wm.txt wm.nii.gz '
my_command+= ' response_gm.txt gm.nii.gz '
my_command+= ' response_csf.txt csf.nii.gz -quiet'
os.system(my_command)
os.remove(subject_results_dir + 'd_img2_selected.nii.gz')
os.remove(subject_results_dir + 'response_wm.txt')
os.remove(subject_results_dir + 'response_gm.txt')
os.remove(subject_results_dir + 'response_csf.txt')
os.remove(subject_results_dir + 'b_vals_selected.bval')
os.remove(subject_results_dir + 'b_vecs_selected.bvec')
# Apply CSD on all target_bval volumes or on data subset
target_bval = 2600
for N_grad, v_grad in zip(N_grad_list, V_grad_list):
DWI_dhcp_norot = dmri_volume[:,:,:,b_vals == target_bval] #[:,:,:,grad_ind_norot].copy()
for i in range(N_grad):
temp= DWI_dhcp_norot[:,:,:,i]/b0_img
temp[compute_mask<0.5]= 0
DWI_dhcp_norot[:,:,:,i]= temp.copy()
DWI_dhcp_norot[DWI_dhcp_norot<0]= 0
DWI_dhcp_norot[DWI_dhcp_norot>1]= 1
DWI_dhcp_norot = nib.Nifti1Image(DWI_dhcp_norot, affine)
nib.save(DWI_dhcp_norot, subject_results_dir + 'd_mri_subset_' + str(N_grad) + '.nii.gz' )
b_vals_0= b_vals[b_vals==0]
b_vecs_0= b_vecs[:,b_vals==0]
b_vals_selected= b_vals[b_vals==target_bval]#[grad_ind_norot]
b_vecs_selected= b_vecs[:,b_vals==target_bval]#[:,grad_ind_norot]
b_vals_selected= np.hstack((b_vals_0,b_vals_selected))
b_vecs_selected= np.hstack((b_vecs_0,b_vecs_selected))
d_img_0= dmri_volume[:,:,:,b_vals==0]
d_img_selected= dmri_volume[:,:,:,b_vals==target_bval]#[:,:,:,grad_ind_norot]
d_img_selected= np.concatenate((d_img_0,d_img_selected), axis=-1)
np.savetxt(subject_results_dir + 'b_vals_selected.txt', b_vals_selected, delimiter=' ')
np.savetxt(subject_results_dir + 'b_vecs_selected.txt', b_vecs_selected, delimiter=' ')
os.rename('b_vals_selected.txt', 'b_vals_selected.bval')
os.rename('b_vecs_selected.txt', 'b_vecs_selected.bvec')
d_img_selected = nib.Nifti1Image(d_img_selected, affine)
nib.save(d_img_selected, subject_results_dir + 'd_img_selected.nii.gz' )
my_command= prefix+'dwi2response tournier d_img_selected.nii.gz -fslgrad b_vecs_selected.bvec b_vals_selected.bval ' + \
' -mask compute_mask_eroded_img.nii.gz response_csd.txt -quiet'
os.system(my_command)
my_command= prefix+'dwi2fod csd d_img_selected.nii.gz response_csd.txt csd_fodf_mrtrix_1000_' \
+ str(N_grad) + '.nii.gz -mask compute_mask_img.nii.gz '\
+ ' -fslgrad b_vecs_selected.bvec b_vals_selected.bval -quiet'
os.system(my_command)
os.remove(subject_results_dir + 'd_img_selected.nii.gz')
os.remove(subject_results_dir + 'response_csd.txt')
os.remove(subject_results_dir + 'b_vals_selected.bval')
os.remove(subject_results_dir + 'b_vecs_selected.bvec')
print("Data generation took (in hours) ", (time.time()-start_time)/3600)