forked from jakeoung/Unet_pytorch
-
Notifications
You must be signed in to change notification settings - Fork 1
/
dataset.py
129 lines (106 loc) · 3.52 KB
/
dataset.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
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
import torch.utils.data as data
import torch
from scipy.ndimage import imread
import os
import os.path
import glob
import numpy as np
from torchvision import transforms
# def make_dataset(root, train=True):
#
# dataset = []
#
# if train:
# dir = os.path.join(root, 'train')
#
# for fGT in glob.glob(os.path.join(dir, '*_mask.tif')):
# fName = os.path.basename(fGT)
# fImg = fName[:-9] + '.tif'
#
# dataset.append( [os.path.join(dir, fImg), os.path.join(dir, fName)] )
#
# return dataset
def make_dataset(root, train=True):
dataset = []
if train:
label_dir = os.path.join(root, 'split_labels')
image_dir = os.path.join(root, 'split_images')
for fGT in glob.glob(os.path.join(label_dir, '*.tif')):
fName = os.path.basename(fGT)
name_str = fName.split('_')
flag_name = '_'+ name_str[len(name_str)-3]+'_'+ name_str[len(name_str)-2] + '_'+name_str[len(name_str)-1]
fImg = glob.glob(os.path.join(image_dir, "*"+flag_name))
if len(fImg) != 1:
assert False
print("Get the image name failed")
fImg = fImg[0]
dataset.append( [fImg, fGT] )
else:
image_dir = os.path.join(root, 'inf_split_images')
dataset = glob.glob(os.path.join(image_dir, '*.tif'))
# for img in glob.glob(os.path.join(image_dir, '*.tif')):
# dataset.append([img])
return dataset
def getImg_count(dir):
files = glob.glob(os.path.join(dir, '*.tif'))
return len(files)
class kaggle2016nerve(data.Dataset):
"""
Read dataset of kaggle ultrasound nerve segmentation dataset
https://www.kaggle.com/c/ultrasound-nerve-segmentation
"""
def __init__(self, root, transform=None, train=True):
self.train = train
self.root = root
# we cropped the image
self.nRow = 480
self.nCol = 480
if self.train:
self.train_set_path = make_dataset(root, train)
else:
self.train_set_path = make_dataset(root, train)
def __getitem__(self, idx):
if self.train:
img_path, gt_path = self.train_set_path[idx]
img = imread(img_path)
#img.resize(self.nRow,self.nCol)
img = img[0:self.nRow, 0:self.nCol]
img = np.atleast_3d(img).transpose(2, 0, 1).astype(np.float32)
if (img.max() - img.min()) < 0.01:
pass
else:
img = (img - img.min()) / (img.max() - img.min())
img = torch.from_numpy(img).float()
gt = imread(gt_path)[0:self.nRow, 0:self.nCol]
gt = np.atleast_3d(gt).transpose(2, 0, 1)
#gt = gt / 255.0 # we don't need to scale
gt = torch.from_numpy(gt).float()
return img, gt
else:
img_path = self.train_set_path[idx]
img_name_noext = os.path.splitext(os.path.basename(img_path))[0]
img = imread(img_path)
# img.resize(self.nRow,self.nCol)
img = img[0:self.nRow, 0:self.nCol]
img = np.atleast_3d(img).transpose(2, 0, 1).astype(np.float32)
if (img.max() - img.min()) < 0.01:
pass
else:
img = (img - img.min()) / (img.max() - img.min())
img = torch.from_numpy(img).float()
return img,img_name_noext
def __len__(self):
if self.train:
# train image count
label_dir = os.path.join(self.root, 'split_labels')
count = getImg_count(label_dir)
print("Image count for training is %d"%count)
return count
# return 5635
# test image count
else:
label_dir = os.path.join(self.root, 'inf_split_images')
count = getImg_count(label_dir)
print("Image count for inference is %d"%count)
return count
# return 5508 # test image count