def test_zscore_normalization(self):
from intensity_normalization.normalize import zscore
normalized = zscore.zscore_normalize(self.img, self.brain_mask)
self.assertAlmostEqual(np.mean(
normalized.get_data()[self.brain_mask.get_data() == 1]),
0,
places=4)
def main():
#util.generate_nifti_images(util.ORIGINAL_DATA) if the data is already in the nifti folder, unnecessary to call again
fols = os.listdir(util.NIFTI_DATA)
i = 1
for folname in fols:
for fname in os.listdir(os.path.join(util.NIFTI_DATA, folname)):
picnift = nib.load(os.path.join(util.NIFTI_DATA, folname, fname))
normalized = zscore.zscore_normalize(
picnift) #this is the most basic normalization
img_data = normalized.get_fdata()
s = img_data.shape
img_data = img_data[:, :, util.SKIP_BOTTOM:(s[2] - util.SKIP_TOP)]
#img_data is type numpy ndarray
s = img_data.shape
mat1 = np.zeros((1, s[0], s[1], s[2]), dtype=np.float32)
if (i == 1):
print(s)
i = i + 1
slice_0 = img_data[int(np.floor(s[0] / 2)), :, :]
slice_1 = img_data[:, int(np.floor(s[1] / 2)), :]
slice_2 = img_data[:, :, int(np.floor(s[2] / 2))]
util.show_slices([slice_0, slice_1, slice_2])
plt.suptitle("Center slices for 1st patient") # doctest: +SKIP
plt.show()
mat1[0] = img_data
filename = folname
mat_dict = {}
mat_dict['data'] = mat1
def process(image_fn, brain_mask_fn, output_dir, logger):
img = io.open_nii(image_fn)
dirname, base, _ = io.split_filename(image_fn)
if output_dir is not None:
dirname = output_dir
if not os.path.exists(dirname):
logger.info('Making output directory: {}'.format(dirname))
os.mkdir(dirname)
if brain_mask_fn is None:
mask = None
else:
if brain_mask_fn == 'nomask':
mask = 'nomask'
else:
mask = io.open_nii(brain_mask_fn)
normalized = zscore.zscore_normalize(img, mask)
outfile = os.path.join(dirname, base + '_zscore.nii.gz')
logger.info('Normalized image saved: {}'.format(outfile))
io.save_nii(normalized, outfile, is_nii=True)
def run_intensity_zscore(infile, outfolder):
from intensity_normalization.normalize import zscore
from intensity_normalization.utilities import io
try:
filename = infile.split(sep)[-1].split(".")[0]
i = io.open_nii(infile)
b_mask = io.open_nii(
join(outfolder, "robex_masks", filename + "_mask.nii.gz"))
if not exists(join(outfolder, "Robex")):
makedirs(join(outfolder, "Robex"))
print('running intensity zscore...')
normalized = zscore.zscore_normalize(i, b_mask)
io.save_nii(normalized, join(outfolder, filename + "_zscore.nii.gz"))
shutil.move(join(infile), join(outfolder, "Robex"))
except:
e = sys.exc_info()
print("Error: ", str(e[0]))
print("Error: ", str(e[1]))
print("Error: executing z-score method")
sys.exit(2)
def get_T1_data(CV=True, fold=0, foldNBR=10, resample=0):
train_MRI = []
test_MRI = []
train_seg = []
test_seg = []
seg = 0
T1 = 0
T2 = 0
T1CE = 0
flair = 0
i = 0
for dirpath, dirnames, filenames in os.walk('D:\BRATS'):
for filename in [
f for f in filenames
if (f.endswith("t1.nii.gz") or f.endswith("t2.nii.gz")
or f.endswith("seg.nii.gz") or f.endswith("t1ce.nii.gz")
or f.endswith("flair.nii.gz"))
]:
img = nib.load(os.path.join(dirpath, filename))
if (i + fold) % foldNBR == 0 and CV:
if filename.endswith("seg.nii.gz"):
#On binarise la segmentation, si on veut segmenter les 4 classes ce n'est pas nécessaire
seg = nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int), img.affine)
if filename.endswith("flair.nii.gz"):
#Le package intensity-normalization fait la normalisation pour nous
flair = zscore.zscore_normalize(
img,
nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int),
img.affine))
if filename.endswith("t1ce.nii.gz"):
T1CE = zscore.zscore_normalize(
img,
nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int),
img.affine))
if filename.endswith("t1.nii.gz"):
T1 = zscore.zscore_normalize(
img,
nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int),
img.affine))
if filename.endswith("t2.nii.gz"):
T2 = zscore.zscore_normalize(
img,
nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int),
img.affine))
#On crop les 5 images de la même façon
flair, T1, T1CE, T2, seg = crop(flair.get_fdata(),
T1.get_fdata(),
T1CE.get_fdata(),
T2.get_fdata(),
seg.get_fdata())
#On reshape pour avoir le format attendu par le network
seg = np.reshape(seg, (155, 194, 148, 1))
T1 = np.reshape(T1, (155, 194, 148, 1))
T2 = np.reshape(T2, (155, 194, 148, 1))
T1CE = np.reshape(T1CE, (155, 194, 148, 1))
flair = np.reshape(flair, (155, 194, 148, 1))
combined = np.concatenate((T1, T2, T1CE, flair), axis=-1)
#IMPORTANT On cast le tout en float16 pour limiter l'espace utilisé
test_seg.append(seg.astype(np.float16))
test_MRI.append(combined.astype(np.float16))
i += 1
else:
if filename.endswith("seg.nii.gz"):
seg = nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int), img.affine)
if filename.endswith("flair.nii.gz"):
flair = zscore.zscore_normalize(
img,
nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int),
img.affine))
if filename.endswith("t1ce.nii.gz"):
T1CE = zscore.zscore_normalize(
img,
nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int),
img.affine))
if filename.endswith("t1.nii.gz"):
T1 = zscore.zscore_normalize(
img,
nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int),
img.affine))
if filename.endswith("t2.nii.gz"):
T2 = zscore.zscore_normalize(
img,
nib.Nifti1Image(
img.get_fdata().astype(bool).astype(int),
img.affine))
flair, T1, T1CE, T2, seg = crop(flair.get_fdata(),
T1.get_fdata(),
T1CE.get_fdata(),
T2.get_fdata(),
seg.get_fdata())
seg = np.reshape(seg, (155, 194, 148, 1))
T1 = np.reshape(T1, (155, 194, 148, 1))
T2 = np.reshape(T2, (155, 194, 148, 1))
T1CE = np.reshape(T1CE, (155, 194, 148, 1))
flair = np.reshape(flair, (155, 194, 148, 1))
combined = np.concatenate((T1, T2, T1CE, flair), axis=-1)
train_seg.append(seg.astype(np.float16))
train_MRI.append(combined.astype(np.float16))
i += 1
return np.array(train_MRI), np.array(train_seg), np.array(
test_MRI), np.array(test_seg)