-
Notifications
You must be signed in to change notification settings - Fork 0
/
dataloader.py
157 lines (115 loc) · 3.72 KB
/
dataloader.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
import pandas as pd
from torch.utils.data import Dataset, DataLoader
import glob
import torch
import cv2
import numpy as np
from scipy.ndimage import zoom
"""
covid ct image the maximum number of cr assumed 70
Parameters:
-t (--text): show the text interface
-h (--help): display this help
"""
''' fix ct number'''
# Resize 2D slices
w, h = 512, 512
def rs_img(img):
'''W and H is 128 now
'''
# print("type",type(img))
img = np.transpose(img)
# print(img.shape,type(img))
flatten = [cv2.resize(img[:,:,i], (w, h), interpolation=cv2.INTER_CUBIC) for i in range(img.shape[-1])]
img = np.array(np.dstack(flatten))
return img
# Spline interpolated zoom (SIZ)
def change_depth_siz(img):
desired_depth = 64
current_depth = img.shape[-1]
depth = current_depth / desired_depth
depth_factor = 1 / depth
img_new = zoom(img, (1, 1, depth_factor), mode='nearest')
return img_new
def normalize(image):
global MIN_BOUND
global MAX_BOUND
image = (image - MIN_BOUND) / (MAX_BOUND - MIN_BOUND)
image[image>1] = 1.
image[image<0] = 0.
return image
def zero_center(image):
image = image - PIXEL_MEAN
return image
def plot_seq(data, name):
a, b = 3, 20
data = np.reshape(data, (a, b, 512, 512))
test_data = data
r, c = test_data.shape[0], test_data.shape[1]
cmaps = [['viridis', 'binary'], ['plasma', 'coolwarm'], ['Greens', 'copper']]
heights = [a[0].shape[0] for a in test_data]
widths = [a.shape[1] for a in test_data[0]]
fig_width = 10. # inches
fig_height = fig_width * sum(heights) / sum(widths)
f, axarr = plt.subplots(r,c, figsize=(fig_width, fig_height),
gridspec_kw={'height_ratios':heights})
for i in range(r):
for j in range(c):
axarr[i, j].imshow(test_data[i][j], cmap='gray')
axarr[i, j].axis('off')
plt.subplots_adjust(wspace=0, hspace=0, left=0, right=1, bottom=0, top=1)
plt.savefig('{}/{}.png'.format('.', name), dpi=1000)
plt.show()
# NOT USING THIS NOW
def resize_depth_wise(img3d):
'''
Inputs a 3d tensor with uneven depth
Outputs a 3d tensor with even depth, in this case depth=64
'''
# patient image 3D
p = img3d
# list of 3D slices of p
p_2d = []
depth = 2
n = 0
c = 0
for c in range(70):
img = img3d[:,:,n+depth]
p_2d.append(img)
n = n+depth
c = c+1
p_3d_d64 = np.array(np.dstack(p_2d))
return p_3d_d64
''' covid class data loader
input: ad
'''
class covid_ct(Dataset,):
def __init__(self, root, csv_file):
self.root = root
self.image_folders = glob.glob('/home/tookai-1/Desktop/sara/covid_safavi/code01/dataset/edited/*')
# print(csv_file)
self.data = pd.read_csv(csv_file).iloc[:, :]
def __len__(self):
return len(self.data )
def __getitem__(self, index):
# print(self.image_folders)
image_name = self.image_folders[index]
image_list = glob.glob(image_name+'/lung_white/*.jpg')
covid_images = []
len_imgs = len(image_list)
v = np.ones([len_imgs,512,512])
i = 0
patiant_id = image_name.split('/')[-1]
for x in image_list:
image1 = cv2.imread(x,0)
img = torch.from_numpy(image1)
v[i,:,:] =img
i = i+1
print(patiant_id)
label = self.data.query('id==@patiant_id')['label'].iloc[0]
features = self.data.query('id==@patiant_id').loc[:,'men':'CRP']
img = rs_img(v)
img_siz = change_depth_siz(img)
img_siz = np.transpose(img_siz)
inputs = np.reshape(img_siz, (64,512,512,1))
return(inputs,label,features.to_numpy())