-
Notifications
You must be signed in to change notification settings - Fork 0
/
SceneFlowDataset.py
88 lines (69 loc) · 2.9 KB
/
SceneFlowDataset.py
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
import os
import numpy as np
from torch.utils.data import Dataset
from torchvision.transforms import ToTensor
from ..utils.utils_io import webp_loader, pfm_loader
class SceneFlowDataset(Dataset):
def __init__(self,
root_path,
transforms,
string_exclude,
string_include
):
self.root_path = root_path
self.string_exclude = string_exclude
self.string_include = string_include
self.transforms = transforms
self.left_image_path, self.right_image_path, \
self.left_disp_path, self.right_disp_path = \
self.get_paths()
def __len__(self):
return len(self.left_image_path)
def __getitem__(self, index: int):
left = webp_loader(self.left_image_path[index])
right = webp_loader(self.right_image_path[index])
disp_left, _ = pfm_loader(self.left_disp_path[index])
disp_left = disp_left[..., np.newaxis]
disp_left = np.ascontiguousarray(disp_left)
disp_right, _ = pfm_loader(self.right_disp_path[index])
disp_right = disp_right[..., np.newaxis]
disp_right = np.ascontiguousarray(disp_right)
sample = {
'left': left, 'right': right,
'disp_left': disp_left, 'disp_right': disp_right
}
torch_sample = ToTensor()(sample)
for transform in self.transforms:
torch_sample = transform(torch_sample)
return torch_sample
def get_paths(self):
left_image_path = []
right_image_path = []
left_disp_path = []
right_disp_path = []
for root, _, files in os.walk(f'{self.root_path}'):
for file in files:
if file.endswith('.webp') and root.endswith(r'\left'):
left_image_item = f'{root}\{file}'
left_image_path.append(left_image_item)
for root, _, files in os.walk(f'{self.root_path}'):
for file in files:
if file.endswith('.webp') and root.endswith(r'\right'):
right_image_item = f'{root}\{file}'
right_image_path.append(right_image_item)
left_disp_path = []
for root, _, files in os.walk(f'{self.root_path}'):
for file in files:
if file.endswith('.pfm') and root.endswith(r'\left'):
left_disp_item = f'{root}\{file}'
left_disp_path.append(left_disp_item)
right_disp_path = []
for root, _, files in os.walk(f'{self.root_path}'):
for file in files:
if file.endswith('.pfm') and root.endswith(r'\right'):
right_disp_item = f'{root}\{file}'
right_disp_path.append(right_disp_item)
return (
left_image_path, right_image_path,
left_disp_path, right_disp_path
)