|
1
|
|
|
from typing import Union, Tuple, Optional, List |
|
2
|
|
|
import torch |
|
3
|
|
|
from ....torchio import DATA |
|
4
|
|
|
from ....data.subject import Subject |
|
5
|
|
|
from ....utils import to_tuple |
|
6
|
|
|
from .. import RandomTransform |
|
7
|
|
|
|
|
8
|
|
|
|
|
9
|
|
|
class RandomFlip(RandomTransform): |
|
10
|
|
|
"""Reverse the order of elements in an image along the given axes. |
|
11
|
|
|
|
|
12
|
|
|
Args: |
|
13
|
|
|
axes: Axis or tuple of axes along which the image will be flipped. |
|
14
|
|
|
flip_probability: Probability that the image will be flipped. This is |
|
15
|
|
|
computed on a per-axis basis. |
|
16
|
|
|
p: Probability that this transform will be applied. |
|
17
|
|
|
seed: See :py:class:`~torchio.transforms.augmentation.RandomTransform`. |
|
18
|
|
|
|
|
19
|
|
|
.. note:: If the input image is 2D, all axes should be in ``(0, 1)``. |
|
20
|
|
|
""" |
|
21
|
|
|
|
|
22
|
|
|
def __init__( |
|
23
|
|
|
self, |
|
24
|
|
|
axes: Union[int, Tuple[int, ...]] = 0, |
|
25
|
|
|
flip_probability: float = 0.5, |
|
26
|
|
|
p: float = 1, |
|
27
|
|
|
seed: Optional[int] = None, |
|
28
|
|
|
): |
|
29
|
|
|
super().__init__(p=p, seed=seed) |
|
30
|
|
|
self.axes = self.parse_axes(axes) |
|
31
|
|
|
self.flip_probability = self.parse_probability( |
|
32
|
|
|
flip_probability, |
|
33
|
|
|
) |
|
34
|
|
|
|
|
35
|
|
|
def apply_transform(self, sample: Subject) -> dict: |
|
36
|
|
|
axes_to_flip_hot = self.get_params(self.axes, self.flip_probability) |
|
37
|
|
|
random_parameters_dict = {'axes': axes_to_flip_hot} |
|
38
|
|
|
items = sample.get_images_dict(intensity_only=False).items() |
|
39
|
|
|
for image_name, image_dict in items: |
|
40
|
|
|
data = image_dict[DATA] |
|
|
|
|
|
|
41
|
|
|
is_2d = data.shape[-3] == 1 |
|
42
|
|
|
dims = [] |
|
43
|
|
|
for dim, flip_this in enumerate(axes_to_flip_hot): |
|
44
|
|
|
if not flip_this: |
|
45
|
|
|
continue |
|
46
|
|
|
actual_dim = dim + 1 # images are 4D |
|
47
|
|
|
# If the user is using 2D images and they use (0, 1) for axes, |
|
48
|
|
|
# they probably mean (1, 2). This should make this transform |
|
49
|
|
|
# more user-friendly. |
|
50
|
|
|
if is_2d: |
|
51
|
|
|
actual_dim += 1 |
|
52
|
|
|
if actual_dim > 3: |
|
53
|
|
|
message = ( |
|
54
|
|
|
f'Image "{image_name}" with shape {data.shape} seems to' |
|
55
|
|
|
' be 2D, so all axes must be in (0, 1),' |
|
56
|
|
|
f' but they are {self.axes}' |
|
57
|
|
|
) |
|
58
|
|
|
raise RuntimeError(message) |
|
59
|
|
|
dims.append(actual_dim) |
|
60
|
|
|
data = torch.flip(data, dims=dims) |
|
61
|
|
|
image_dict[DATA] = data |
|
62
|
|
|
sample.add_transform(self, random_parameters_dict) |
|
63
|
|
|
return sample |
|
64
|
|
|
|
|
65
|
|
|
@staticmethod |
|
66
|
|
|
def get_params(axes: Tuple[int, ...], probability: float) -> List[bool]: |
|
67
|
|
|
axes_hot = [False, False, False] |
|
68
|
|
|
for axis in axes: |
|
69
|
|
|
random_number = torch.rand(1) |
|
70
|
|
|
flip_this = bool(probability > random_number) |
|
71
|
|
|
axes_hot[axis] = flip_this |
|
72
|
|
|
return axes_hot |
|
73
|
|
|
|
|
74
|
|
|
@staticmethod |
|
75
|
|
|
def parse_axes(axes: Union[int, Tuple[int, ...]]): |
|
76
|
|
|
axes_tuple = to_tuple(axes) |
|
77
|
|
|
for axis in axes_tuple: |
|
78
|
|
|
is_int = isinstance(axis, int) |
|
79
|
|
|
if not is_int or axis not in (0, 1, 2): |
|
80
|
|
|
raise ValueError('All axes must be 0, 1 or 2') |
|
81
|
|
|
return axes_tuple |
|
82
|
|
|
|
fepegar /
torchio
Loading history...