Download this file

176 lines (134 with data), 6.0 kB

  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
import os
import re
import numpy as np
import SimpleITK as sitk
from skimage import filters
import seaborn as sns
import matplotlib.pyplot as plt
from matplotlib.lines import Line2D
def resample_img(img,
target_direction, new_origin,
target_voxel_spacing, target_shape,
default_value, interpolator):
transformation = sitk.ResampleImageFilter()
transformation.SetOutputDirection(target_direction)
transformation.SetOutputOrigin(new_origin)
transformation.SetOutputSpacing(target_voxel_spacing)
transformation.SetSize(target_shape)
transformation.SetDefaultPixelValue(default_value)
transformation.SetInterpolator(interpolator)
return transformation.Execute(img)
def normalize_img(img, window_min, window_max):
"""
Transform input value from window_min - window_max to 0 - 1
"""
intensityWindowingFilter = sitk.IntensityWindowingImageFilter()
intensityWindowingFilter.SetOutputMaximum(1)
intensityWindowingFilter.SetOutputMinimum(0)
intensityWindowingFilter.SetWindowMaximum(window_max)
intensityWindowingFilter.SetWindowMinimum(window_min)
return intensityWindowingFilter.Execute(img)
def normalize_img_v2(img, shift, scale):
return sitk.ShiftScale(img, shift=shift, scale=scale)
def threshold_img(img, threshold):
return sitk.Threshold(img, lower=0.0, upper=threshold, outsideValue=threshold)
def mip(img, threshold=None):
img_array = sitk.GetArrayFromImage(img)
if threshold:
# img_array = np.array(img_array>threshold, dtype=np.int8)
img_array[img_array > threshold] = threshold
return np.max(img_array, axis=1), np.max(img_array, axis=2)
def get_info(img):
print('img information :')
print('\t Origin :', img.GetOrigin())
print('\t Size :', img.GetSize())
print('\t Spacing :', img.GetSpacing())
print('\t Direction :', img.GetDirection())
def get_study_uid(img_path):
return re.sub('_nifti_(PT|mask|CT)\.nii(\.gz)?', '', os.path.basename(img_path))
def one_hot_encode(x, n_classes=None):
"""
One hot encode a list of sample labels. Return a one-hot encoded vector for each label.
: x: List of sample Labels
: return: Numpy array of one-hot encoded labels
"""
if n_classes is None:
n_classes = np.max(x) + 1
return np.eye(n_classes)[x]
def roi2tmtv(mask_img, pet_img, threshold='auto'):
"""
Generate the mask from the ROI of the pet scan
Args:
:param mask_img: sitk image, raw mask (i.e ROI)
:param pet_img: sitk image, the corresponding pet scan
:param threshold: threshold to apply to the ROI to get the tumor segmentation.
if set to 'auto', it will take 42% of the maximum
:return: sitk image, the ground truth segmentation
"""
# transform to numpy
mask_array = sitk.GetArrayFromImage(mask_img)
pet_array = sitk.GetArrayFromImage(pet_img)
# get 3D meta information
if len(mask_array.shape) == 3:
mask_array = np.expand_dims(mask_array, axis=0)
origin = mask_img.GetOrigin()
spacing = mask_img.GetSpacing()
direction = tuple(mask_img.GetDirection())
size = mask_img.GetSize()
else:
origin = mask_img.GetOrigin()[:-1]
spacing = mask_img.GetSpacing()[:-1]
direction = tuple(el for i, el in enumerate(mask_img.GetDirection()[:12]) if not (i + 1) % 4 == 0)
size = mask_img.GetSize()[:-1]
# print(pet_img.GetOrigin(), origin)
# print(pet_img.GetSpacing(), spacing)
# print(pet_img.GetDirection(), direction)
# print(pet_img.GetSize(), size)
# print(mask_array.shape)
# assert meta-info roi == meta-info pet
assert pet_img.GetOrigin() == origin
assert pet_img.GetSpacing() == spacing
assert pet_img.GetDirection() == direction
assert pet_img.GetSize() == size
# generate mask from ROIs
new_mask = np.zeros(mask_array.shape[1:], dtype=np.int8)
n_voxels_per_roi = np.zeros(mask_array.shape[0])
for num_slice in range(mask_array.shape[0]):
mask_slice = mask_array[num_slice]
# calculate threshold value of the roi
if threshold == 'auto':
roi = pet_array[mask_slice > 0]
if len(roi) > 0:
SUV_max = np.max(roi)
threshold_suv = SUV_max * 0.41
else:
threshold_suv = 0.0
elif threshold == 'otsu':
roi = pet_array[mask_slice > 0]
if len(roi) > 0:
threshold_suv = filters.threshold_otsu(roi)
else:
threshold_suv = 0.0
else:
threshold_suv = threshold
# apply threshold
n_voxels_per_roi[num_slice] = np.sum((pet_array >= threshold_suv) & (mask_slice > 0))
new_mask[np.where((pet_array >= threshold_suv) & (mask_slice > 0))] = 1
return np.sum(new_mask), n_voxels_per_roi
def plot_hist_roi(mask_img, pet_img):
mask_array = sitk.GetArrayFromImage(mask_img)
pet_array = sitk.GetArrayFromImage(pet_img)
for num_slice in range(mask_array.shape[0]):
mask_slice = mask_array[num_slice]
roi = pet_array[mask_slice > 0]
sns.distplot(roi, hist=True, kde=True, rug=True, rug_kws={'color': 'black'})
# colors = [cmap(0.), cmap(0.5), cmap(1.0)] # ['red', 'green', 'orange']
colors = ['red', 'green', 'orange']
for threshold_val, color in zip([filters.threshold_otsu(roi), 2.5, 0.42 * np.max(roi)], colors):
plt.plot([threshold_val, threshold_val], [0.0, 1.0], '--', color=color)
custom_lines = [Line2D([0], [0], color=colors[0], lw=4),
Line2D([0], [0], color=colors[1], lw=4),
Line2D([0], [0], color=colors[2], lw=4)]
plt.legend(custom_lines, ['otsu', '2.5 SUV', '42%'])
# plt.savefig('plot/train{}_slice{}_hist'.format(idx, num_slice))
plt.show()