Multi-Modal-Brain-Segmentation/normalize_data.py at master · shivangi-aneja/Multi-Modal-Brain-Segmentation · GitHub

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
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""

@author: Shivangi
"""

from __future__ import print_function
import scipy
import os
import numpy as np
import SimpleITK as sitk
from scipy import ndimage
import nibabel as nib

train_list = ['1', '4', '5', '70']
val_list = ['148']
test_list = ['7', '14']
unlabelled_list = ['0','2','4','6','8','11','17','19','21','23','25']
cut = 10
thresh = 10


def normalize_annotated_images(data_dir, dir_list, dest_dir):
    """

    :param dir_list:
    :return:
    """
    for dir_name in dir_list:

        # Read the segmentation mask and cut the borders
        seg_img = sitk.ReadImage(os.path.join(data_dir, dir_name, 'segm.nii.gz'))
        seg_array = sitk.GetArrayFromImage(seg_img)
        seg_array = seg_array[:, cut:np.shape(seg_array)[1] - cut, cut:np.shape(seg_array)[2] - cut]

        # Read the FLAIR image and cut the borders
        flair_img = sitk.ReadImage(os.path.join(data_dir, dir_name, 'FLAIR.nii.gz'))
        flair_array = sitk.GetArrayFromImage(flair_img)

        # Read the T1 image and cut the borders
        t1_img = sitk.ReadImage(os.path.join(data_dir, dir_name, 'reg_T1.nii.gz'))
        t1_array = sitk.GetArrayFromImage(t1_img)


        # Normalize the cut FLAIR image
        brain_mask_flair = np.zeros(np.shape(flair_array), dtype='float32')
        brain_mask_flair[flair_array >= thresh] = 1
        brain_mask_flair[flair_array < thresh] = 0
        for iii in range(np.shape(flair_array)[0]):
            brain_mask_flair[iii, :, :] = scipy.ndimage.morphology.binary_fill_holes(
                brain_mask_flair[iii, :, :])  # fill the holes inside brain
        flair_array = flair_array - np.mean(flair_array[brain_mask_flair == 1])
        flair_array /= np.std(flair_array[brain_mask_flair == 1])
        flair_array = flair_array[:, cut:np.shape(flair_array)[1] - cut, cut:np.shape(flair_array)[2] - cut]


        # Normalize the cut regualrized T1 image
        brain_mask_t1 = np.zeros(np.shape(t1_array), dtype='float32')
        brain_mask_t1[t1_array >= thresh] = 1
        brain_mask_t1[t1_array < thresh] = 0
        for iii in range(np.shape(t1_array)[0]):
            brain_mask_t1[iii, :, :] = scipy.ndimage.morphology.binary_fill_holes(
                brain_mask_t1[iii, :, :])  # fill the holes inside br
        t1_array = t1_array - np.mean(t1_array[brain_mask_t1 == 1])
        t1_array /= np.std(t1_array[brain_mask_t1 == 1])
        t1_array = t1_array[:, cut:np.shape(t1_array)[1] - cut, cut:np.shape(t1_array)[2] - cut]

        if not os.path.exists(os.path.join(dest_dir, dir_name)):
            os.makedirs(os.path.join(dest_dir, dir_name))

        # Save the segmentation mask
        seg_image = nib.Nifti1Image(np.transpose(seg_array,[1,2,0]), None)
        imgname = 'segm.nii.gz'
        nib.save(seg_image, os.path.join(os.path.join(dest_dir, dir_name), imgname))

        # Save the FLAIR image
        flair_image = nib.Nifti1Image(np.transpose(flair_array,[1,2,0]), None)
        imgname = 'FLAIR.nii.gz'
        nib.save(flair_image, os.path.join(os.path.join(dest_dir, dir_name), imgname))

        # Save the T1 image
        t1_image = nib.Nifti1Image(np.transpose(t1_array,[1,2,0]), None)
        imgname = 'reg_T1.nii.gz'
        nib.save(t1_image, os.path.join(os.path.join(dest_dir, dir_name), imgname))


def normalize_unannotated_images(data_dir, dir_list, dest_dir):
    """

    :param dir_list:
    :return:
    """
    for dir_name in dir_list:

        print("Start " + os.path.join(data_dir, dir_name))
        # Read the FLAIR image and cut the borders
        flair_img = sitk.ReadImage(os.path.join(data_dir, dir_name, 'FLAIR.nii.gz'))
        flair_array = sitk.GetArrayFromImage(flair_img)

        # Read the T1 image and cut the borders
        t1_img = sitk.ReadImage(os.path.join(data_dir, dir_name, 'T1.nii.gz'))
        t1_array = sitk.GetArrayFromImage(t1_img)


        # Normalize the cut FLAIR image
        brain_mask_flair = np.zeros(np.shape(flair_array), dtype='float32')
        brain_mask_flair[flair_array >= thresh] = 1
        brain_mask_flair[flair_array < thresh] = 0
        for iii in range(np.shape(flair_array)[0]):
            brain_mask_flair[iii, :, :] = scipy.ndimage.morphology.binary_fill_holes(
                brain_mask_flair[iii, :, :])  # fill the holes inside brain
        flair_array = flair_array - np.mean(flair_array[brain_mask_flair == 1])
        flair_array /= np.std(flair_array[brain_mask_flair == 1])
        flair_array = flair_array[:, cut:np.shape(flair_array)[1] - cut, cut:np.shape(flair_array)[2] - cut]


        # Normalize the cut regualrized T1 image
        brain_mask_t1 = np.zeros(np.shape(t1_array), dtype='float32')
        brain_mask_t1[t1_array >= thresh] = 1
        brain_mask_t1[t1_array < thresh] = 0
        for iii in range(np.shape(t1_array)[0]):
            brain_mask_t1[iii, :, :] = scipy.ndimage.morphology.binary_fill_holes(
                brain_mask_t1[iii, :, :])  # fill the holes inside br
        t1_array = t1_array - np.mean(t1_array[brain_mask_t1 == 1])
        t1_array /= np.std(t1_array[brain_mask_t1 == 1])
        t1_array = t1_array[:, cut:np.shape(t1_array)[1] - cut, cut:np.shape(t1_array)[2] - cut]

        if not os.path.exists(os.path.join(dest_dir, dir_name)):
            os.makedirs(os.path.join(dest_dir, dir_name))

        # Save the FLAIR image
        flair_image = nib.Nifti1Image(np.transpose(flair_array,[1,2,0]), None)
        imgname = 'FLAIR.nii.gz'
        nib.save(flair_image, os.path.join(os.path.join(dest_dir, dir_name), imgname))

        # Save the T1 image
        t1_image = nib.Nifti1Image(np.transpose(t1_array,[1,2,0]), None)
        imgname = 'T1.nii.gz'
        nib.save(t1_image, os.path.join(os.path.join(dest_dir, dir_name), imgname))
        print("End " + os.path.join(data_dir, dir_name))


if __name__ == '__main__':

    # Normalize training data
    normalize_annotated_images(data_dir=os.getcwd() + "/data/mrbrains/train/", dir_list=train_list,
                               dest_dir=os.getcwd()+"/data/mrbrains_normalized/train/")
    # Normalize validation data
    normalize_annotated_images(data_dir=os.getcwd() + "/data/mrbrains/val/", dir_list=val_list,
                               dest_dir=os.getcwd() + "/data/mrbrains_normalized/val/")
    # Normalize test data
    normalize_annotated_images(data_dir=os.getcwd() + "/data/mrbrains/test/", dir_list=test_list,
                               dest_dir=os.getcwd() + "/data/mrbrains_normalized/test/")

    # Normalize unlabelled data
    normalize_unannotated_images(data_dir=os.getcwd() + "/data/mrbrains/unlabelled/", dir_list=unlabelled_list,
                               dest_dir=os.getcwd() + "/data/mrbrains_normalized/unlabelled/")