-
Notifications
You must be signed in to change notification settings - Fork 1
/
skullSSM.py
155 lines (109 loc) · 4.18 KB
/
skullSSM.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
import nrrd
import gc
import ants, shutil
from glob import glob
from sklearn.decomposition import PCA
class skullRecSSM(object):
def __init__(self,numOfImg4SSM=30):
self.numOfImg4SSM=numOfImg4SSM
def reg(self,fix_img,moving_img):
outs = ants.registration(fix_img, moving_img, type_of_transforme = 'Similarity')
warped_img = outs['warpedmovout']
return warped_img
def inverse_reg(self,fixed1,moving1,moving2):
outs1 = ants.registration(fixed1, moving1, type_of_transforme = 'Similarity')
outs2 = ants.apply_transforms(moving1, moving2,transformlist=outs1['invtransforms'])
outs2=outs2.numpy()
outs2=(outs2>0)
outs2=outs2+1-1
return outs2
def ssm_train(self,warped_train_dir):
complete = glob(warped_train_dir+'*.nrrd')
pca = PCA(n_components=len(complete))
#(512,512,222) is the size of the fixed image
data=np.zeros(shape=(len(complete),512,512,222),dtype='int16')
for i in range(len(complete)):
temp,header=nrrd.read(complete[i])
data[i,:,:,:]=temp
del temp
data=data[0:self.numOfImg4SSM]
self.mean_shape=data.mean(axis=0)
data=np.reshape(data,(len(complete),512*512*222))
data_pca = pca.fit_transform(data)
#explained_variance_ratio
#percentage=pca.explained_variance_ratio_
#components_
#components_=pca.components_
data_inv=np.linalg.pinv(data)
del data
gc.collect()
self.eigenvec=data_inv.dot(data_pca)
del data_
del data_inv
gc.collect()
def ssm_test(self, testImg, useOnlyMeanShape=False):
testdata=np.reshape(testImg,(1,512*512*222))
testdatapca=testdata.dot(self.eigenvec)
lambda_n=[]
for i in range(len(testdatapca)):
lambda_n.append(testdatapca[i])
lambda_n=np.array(lambda_n)
#scale [0,1]
lambda_n = (lambda_n - np. min(lambda_n))/np. ptp(lambda_n)
lambda_n=np.transpose(lambda_n)
reconstructed=eigenvec.dot(lambda_n)
reconstructed=np.reshape(reconstructed,(512,512,222))
if useOnlyMeanShape:
rec=self.mean_shape
else:
rec=reconstructed+self.mean_shape
rec=(rec>0)
rec=rec+1-1
return rec
if __name__ == "__main__":
fixed_img=ants.image_read('./fixed/001.nrrd')
moving_train_dir='./moving_train/'
moving_test_dir='./moving_test/'
warped_train_dir='./warped_img/train/'
warped_test_dir='./warped_img/test/'
results_dir='./results/'
moving_train_imgs = glob(moving_train_dir+'*.nrrd')
moving_test_imgs = glob(moving_test_dir+'*.nrrd')
warped_test_img = glob(warped_test_dir+'*.nrrd')
model=skullRecSSM(30)
print('warpping training images...')
for idx in range(len(moving_train_imgs)):
NamePrefix = str(idx).zfill(3)
moving_img = ants.image_read(moving_train_imgs[idx])
outs = model.reg(fixed_img, moving_img)
warped_img = outs['warpedmovout']
ants.image_write(warped_img, warped_train_dir + NamePrefix +'.nrrd')
rint('building SSM...')
model.ssm_train(warped_train_dir)
print('warpping test images...')
for idx in range(len(moving_test_imgs)):
NamePrefix = str(idx).zfill(3)
moving_img = ants.image_read(moving_test_imgs[idx])
outs = ssm.reg(fixed_img, moving_img)
warped_img = outs['warpedmovout']
ants.image_write(warped_img, warped_test_dir + NamePrefix +'.nrrd')
print('fitting...')
for i in range(len(warped_test_img)):
NamePrefix = str(i).zfill(3)
test,h=nrrd.read(warped_test_img[i])
h['type']='int32'
h['encoding']='gzip'
rec=model.ssm_test(test,useOnlyMeanShape=False)
implant=rec-test
nrrd.write(results_dir+'skulls/'+NamePrefix,rec.astype('int32'),h)
nrrd.write(results_dir+'implants/'+NamePrefix,implant.astype('int32'),h)
print('converting the results back to original image space...')
skull_imgs = glob(results_dir+'skulls/'+'*.nrrd')
implant_imgs = glob(results_dir+'implants/'+'*.nrrd')
for i in range(len(warped_test_img)):
NamePrefix = str(i).zfill(3)
moving = ants.image_read(moving_test_imgs[i])
moving_results = ants.image_read(implant_imgs[i])
converted_img=model.inverse_reg(fixed_img,moving,moving_results)
_,header=nrrd.read(moving_test_imgs[i])
nrrd.write(implant_imgs,converted_img.astype('int32'),header)