def read_file(self):
#print(self.dir_name)
#print(self.file_index)
#print(len(self.image_fname))
print(self.image_fname[self.random_batch[self.file_index]])
print(self.dir_name + '/' + self.dir_image + '/' + self.image_fname[self.random_batch[self.file_index]])
image = m2n.loadimage(self.dir_name + '/' + self.dir_image + '/' + self.image_fname[self.random_batch[self.file_index]])
_, mask, str_name = m2n.parsemask(self.dir_name + '/' + self.dir_mask + '/' + self.mask_fname[self.random_batch[self.file_index]])
mask = mask[self.mask_index[self.random_batch[self.file_index]]]
# Broadcasting from int32 to float64/32(?) is mandatory for resizing images without pixel value change!
image = np.float64(image)
mask = np.float64(mask)
print(str_name[self.mask_index[self.random_batch[self.file_index]]])
self.file_index += 1
if self.opt_resize == True and self.opt_crop == True:
print("Both resize and crop are selected. Please choose either one")
# resize 2-D N number of image (N x width x height)
if self.opt_resize == True:
print("image is resized")
image = dpp.resize_3d(dcmimage=image, resize_shape=self.resize_shape)
mask = dpp.resize_3d(dcmimage=mask, resize_shape=self.resize_shape)
# crop 2-D N number of image (N x width x height)
if self.opt_crop == True:
print("image is cropped")
image = dpp.crop_3d(dcmimage=image, crop_shape=self.crop_shape)
mask = dpp.crop_3d(dcmimage=mask, crop_shape=self.crop_shape)
return image, mask
def read_file(self):
#print(self.dir_name)
#print(self.file_index)
#print(len(self.image_fname))
image = m2n.loadimage(self.dir_name + '/image/' +
self.image_fname[self.file_index])
_, mask, str_name = m2n.parsemask(self.dir_name + '/mask/' +
self.mask_fname[self.file_index])
mask = mask[self.mask_index[self.file_index]]
# Broadcasting from int32 to float64/32(?) is mandatory for resizing images without pixel value change!
image = np.float64(image)
mask = np.float64(mask)
print(str_name[self.mask_index[self.file_index]])
self.file_index += 1
# resize 2-D N number of image (N x width x height)
image = dpp.resize_3d(dcmimage=image, resize_shape=self.resize_shape)
mask = dpp.resize_3d(dcmimage=mask, resize_shape=self.resize_shape)
return image, mask
def __init__(self, dir_name='DICOM_data/training_set', dir_image='image', dir_mask='mask',contour_name='mandible', zero_slices=False, opt_resize=True,
resize_shape=(224,224), opt_crop=False, crop_shape=(224,224),
rotation=True, rotation_angle=[90], bitsampling=True, bitsampling_bit=[4]):
self.voxel_range = np.array([0,3000]) # any voxel bigger than 3000 would be 3000, less than 0 would be 0
self.normalization_range = np.array([0,255])
self.dir_name = dir_name
self.zero_slices = zero_slices # Not taking account of the slices with no mask pixel
self.file_index = 0
self.slice_index = 0
self.opt_resize = opt_resize
self.resize_shape = resize_shape # for VGG16 and VGG19. for inception, (299,299) used
self.opt_crop = opt_crop
self.crop_shape = crop_shape
self.num_channel = 3 # 3 channel input
# Data Augmentation options
self.rotation = rotation
self.rotation_angle = rotation_angle
self.bitsampling = bitsampling
self.bitsampling_bit = bitsampling_bit
self.dir_image = dir_image
self.dir_mask = dir_mask
mask_fname = []
mask_index = []
# Only choose the dataset with given contour name & exclude the
for file in glob.glob(self.dir_name + '/' + self.dir_mask + '/*.mat'):
num_str, mask, name = m2n.parsemask(file)
for i in range(num_str):
if contour_name in name[i][0].lower():
mask_index.append(i)
mask_fname.append(os.path.basename(file))
break
# Matching mask with img
image_fname = [0]*len(mask_fname)
for file in glob.glob(self.dir_name + '/' + self.dir_image + "/*.mat"):
img_name = os.path.basename(file)
for masks in mask_fname:
if re.split("[._]", img_name)[1] == re.split("[._]",masks)[1]:
idx = mask_fname.index(masks)
image_fname[idx] = img_name
# To check if mask & image matched in order
# for i in range(len(mask_fname)):
# print(image_fname[i], mask_fname[i], mask_index[i])
self.image_fname = image_fname
self.mask_fname = mask_fname
self.mask_index = mask_index
self.num_files = len(mask_fname)
# Randomize the order of data
print("Num files : ", self.num_files)
self.random_batch = random.sample(range(0,self.num_files),self.num_files)
print("random_batch: ", self.random_batch)
self.image, self.mask = self.read_file()
def __init__(self,
dir_name='DICOM_data/training_set',
contour_name='mandible',
zero_slices=False,
resize_shape=(227, 227),
rotation=True,
rotation_angle=[90],
bitsampling=True,
bitsampling_bit=[4]):
self.voxel_range = np.array(
[-1000, 1000]
) # any voxel bigger than 1000 would be 1000, less than -1000 would be -1000
self.normalization_range = np.array([0, 255])
self.dir_name = dir_name
self.contour_name = contour_name
self.zero_slices = zero_slices # Not taking account of the slices with no mask pixel
self.file_index = 0
self.slice_index = 0
self.resize_shape = resize_shape # for VGG16 and VGG19. for inception, (299,299) used
self.num_channel = 3 # 3 channel input
# Data Augmentation options
self.rotation = rotation
self.rotation_angle = rotation_angle
self.bitsampling = bitsampling
self.bitsampling_bit = bitsampling_bit
mask_fname = []
mask_index = []
# Only choose the dataset with given contour name
for file in glob.glob(self.dir_name + "/mask/*.mat"):
num_str, mask, name = m2n.parsemask(file)
for i in range(num_str):
if name[i][0].lower() == self.contour_name:
mask_index.append(i)
mask_fname.append(os.path.basename(file))
break
# Matching mask with img
image_fname = [0] * len(mask_fname)
for file in glob.glob(self.dir_name + "/image/*.mat"):
img_name = os.path.basename(file)
for masks in mask_fname:
if re.split("[._]", img_name)[1] == re.split("[._]", masks)[1]:
idx = mask_fname.index(masks)
image_fname[idx] = img_name
# To check if mask & image matched in order
# for i in range(len(mask_fname)):
# print(image_fname[i], mask_fname[i], mask_index[i])
self.image_fname = image_fname
self.mask_fname = mask_fname
self.mask_index = mask_index
self.num_files = len(mask_fname)
self.image, self.mask = self.read_file()