-
Notifications
You must be signed in to change notification settings - Fork 21
/
isic2018.py
executable file
·151 lines (120 loc) · 4.99 KB
/
isic2018.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
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
import os
import glob
import json
import torch
import random
import torch.nn as nn
import numpy as np
import torch.utils.data
from torchvision import transforms
import torch.utils.data as data
import torch.nn.functional as F
import albumentations as A
from sklearn.model_selection import KFold
def norm01(x):
return np.clip(x, 0, 255) / 255
seperable_indexes = json.load(open('dataset/data_split.json', 'r'))
# cross validation
class myDataset(data.Dataset):
def __init__(self, fold, split, aug=False):
super(myDataset, self).__init__()
self.split = split
root_data_dir = '/raid/wjc/data/skin_lesion/isic2018/'
# load images, label, point
self.image_paths = []
self.label_paths = []
self.point_paths = []
self.dist_paths = []
indexes = [l[:-4] for l in os.listdir(root_data_dir + 'Image/')]
valid_indexes = seperable_indexes[fold]
train_indexes = list(filter(lambda x: x not in valid_indexes, indexes))
print('Fold {}: train: {} valid: {}'.format(fold, len(train_indexes),
len(valid_indexes)))
indexes = train_indexes if split == 'train' else valid_indexes
self.image_paths = [
root_data_dir + '/Image/{}.npy'.format(_id) for _id in indexes
]
self.label_paths = [
root_data_dir + '/Label/{}.npy'.format(_id) for _id in indexes
]
self.point_paths = [
root_data_dir + '/Point/{}.npy'.format(_id) for _id in indexes
]
# self.point_All_paths = [
# '/data2/cf_data/skinlesion_segment/ISIC2018_rawdata/ISBI_2018/Train/Point_All/{}.npy'
# .format(_id) for _id in indexes
# ]
print('Loaded {} frames'.format(len(self.image_paths)))
self.num_samples = len(self.image_paths)
self.aug = aug
p = 0.5
self.transf = A.Compose([
A.GaussNoise(p=p),
A.HorizontalFlip(p=p),
A.VerticalFlip(p=p),
A.ShiftScaleRotate(p=p),
# A.RandomBrightnessContrast(p=p),
])
def __getitem__(self, index):
image_data = np.load(self.image_paths[index])
label_data = np.load(self.label_paths[index]) > 0.5
point_data = np.load(self.point_paths[index]) > 0.5
if self.aug and self.split == 'train':
mask = np.concatenate([
label_data[..., np.newaxis].astype('uint8'),
point_data[..., np.newaxis]
],
axis=-1)
# print(mask.shape)
tsf = self.transf(image=image_data.astype('uint8'), mask=mask)
image_data, mask_aug = tsf['image'], tsf['mask']
label_data = mask_aug[:, :, 0]
point_data = mask_aug[:, :, 1]
image_data = norm01(image_data)
label_data = np.expand_dims(label_data, 0)
point_data = np.expand_dims(point_data, 0)
image_data = torch.from_numpy(image_data).float()
label_data = torch.from_numpy(label_data).float()
point_data = torch.from_numpy(point_data).float()
image_data = image_data.permute(2, 0, 1)
return {
'image_path': self.image_paths[index],
'label_path': self.label_paths[index],
'point_path': self.point_paths[index],
'image': image_data,
'label': label_data,
'point': point_data
}
def __len__(self):
return self.num_samples
def dataset_kfold(dataset_dir, save_path, k=5):
indexes = [l[:-4] for l in os.listdir(dataset_dir)]
kf = KFold(k, shuffle=True) #k折交叉验证
val_index = dict()
for i in range(k):
val_index[str(i)] = []
for i, (tr, val) in enumerate(kf.split(indexes)):
for item in val:
val_index[str(i)].append(indexes[item])
print('fold:{},train_len:{},val_len:{}'.format(i, len(tr), len(val)))
with open(save_path, 'w') as f:
json.dump(val_index, f)
def random_seperate_dataset():
# indexes = [l[:-4] for l in os.listdir('/raid/wl/ISBI_2018/Train/Image/')]
# random.shuffle(indexes)
# names = {'0':indexes[:500], '1':indexes[500:1000], '2':indexes[1000:1500], '3':indexes[1500:2000],'4':indexes[2000:]}
# names = [indexes[:500], indexes[500:1000], indexes[1000:1500], indexes[1500:2000], indexes[2000:]]
# with open('/raid/wl/ISBI_2018/data_split.json','w') as f:
# json.dump(names, f)
return
if __name__ == '__main__':
from tqdm import tqdm
dataset = myDataset(fold='0', split='train', aug=True)
train_loader = torch.utils.data.DataLoader(dataset,
batch_size=8,
shuffle=False,
num_workers=2,
pin_memory=True,
drop_last=True)
for d in train_loader:
pass