|
1
|
|
|
from typing import Union, Sequence, Tuple |
|
2
|
|
|
|
|
3
|
|
|
import torch |
|
4
|
|
|
import numpy as np |
|
5
|
|
|
|
|
6
|
|
|
from ...utils import to_tuple |
|
7
|
|
|
|
|
8
|
|
|
|
|
9
|
|
|
class PatchSampler: |
|
10
|
|
|
r"""Base class for TorchIO samplers. |
|
11
|
|
|
|
|
12
|
|
|
Args: |
|
13
|
|
|
patch_size: Tuple of integers :math:`(d, h, w)` to generate patches |
|
14
|
|
|
of size :math:`d \times h \times w`. |
|
15
|
|
|
If a single number :math:`n` is provided, :math:`d = h = w = n`. |
|
16
|
|
|
""" |
|
17
|
|
|
def __init__(self, patch_size: Union[int, Sequence[int]]): |
|
18
|
|
|
patch_size = np.array(to_tuple(patch_size, length=3)) |
|
19
|
|
|
if np.any(patch_size < 1): |
|
20
|
|
|
message = ( |
|
21
|
|
|
'Patch dimensions must be positive integers,' |
|
22
|
|
|
f' not {patch_size}' |
|
23
|
|
|
) |
|
24
|
|
|
raise ValueError(message) |
|
25
|
|
|
self.patch_size = patch_size.astype(np.uint16) |
|
26
|
|
|
|
|
27
|
|
|
def __call__(self): |
|
28
|
|
|
raise NotImplementedError |
|
29
|
|
|
|
|
30
|
|
|
def get_probability_map(self): |
|
31
|
|
|
raise NotImplementedError |
|
32
|
|
|
|
|
33
|
|
|
def extract_patch(self): |
|
34
|
|
|
raise NotImplementedError |
|
35
|
|
|
|
|
36
|
|
|
@staticmethod |
|
37
|
|
|
def get_crop_transform(sample, index_ini, patch_size): |
|
38
|
|
|
from ...transforms.preprocessing.spatial.crop import Crop |
|
39
|
|
|
shape = np.array(sample.spatial_shape, dtype=np.uint16) |
|
40
|
|
|
index_ini = np.array(index_ini, dtype=np.uint16) |
|
41
|
|
|
patch_size = np.array(patch_size, dtype=np.uint16) |
|
42
|
|
|
index_fin = index_ini + patch_size |
|
43
|
|
|
crop_ini = index_ini.tolist() |
|
44
|
|
|
crop_fin = (shape - index_fin).tolist() |
|
45
|
|
|
cropping = sum(zip(crop_ini, crop_fin), start=()) |
|
46
|
|
|
return Crop(cropping) |
|
47
|
|
|
|
fepegar /
torchio
