-
Notifications
You must be signed in to change notification settings - Fork 0
/
normalize.py
76 lines (57 loc) · 2.38 KB
/
normalize.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
import torch
from torch.utils.data import Dataset, Subset, DataLoader, random_split
import os
import numpy as np
import matplotlib.pyplot as plt
import math
# TODO: Construct your data in the following baseline structure: 1) ./Dataset/Train/image/, 2) ./Dataset/Train/label, 3) ./Dataset/Test/image, and 4) ./Dataset/Test/label
class LungDataset(Dataset):
def __init__(self, root):
self.root = root
def __len__(self):
# Return number of points in the dataset based on root path
imgs_path = os.path.join(self.root, 'image')
return len(os.listdir(imgs_path))
def __getitem__(self, idx):
# Here we have to return the item requested by `idx`. The PyTorch DataLoader class will use this method to make an iterable for training/validation loop.
# File names are based on idx.
img_path = os.path.join(self.root, 'image', f'{str(idx)}.png')
label_path = os.path.join(self.root, 'label', f'{str(idx)}.txt')
# Import image
# Transpose to be 3x244x244
image = np.transpose(torch.tensor(plt.imread(img_path)), (2, 0, 1))
# Normalize image to reduce computation
# image = transforms.Normalize(parameters['mean'], parameters['std']).forward(image)\
# Greyscale image
# image = transforms.Grayscale(num_output_channels=3).forward(image)
# Get label of corresponding image
l = open(label_path, 'r')
label = int(l.read())
# Return manipulated image and label
return image, label
fold = [
LungDataset('./Dataset/fold0'),
LungDataset('./Dataset/fold1'),
LungDataset('./Dataset/fold2'),
LungDataset('./Dataset/fold3'),
LungDataset('./Dataset/fold4')
]
mean = torch.tensor(0)
for i in range(5):
# Copy fold array
training = fold.copy()
testing = DataLoader(training[i], batch_size=32, shuffle=True)
del training[i]
training_count = 0
for dataset in range(0, len(training)):
training_count = training_count + len(training[dataset])
print(training_count)
mean = torch.tensor(0)
std = torch.tensor(0)
for dataset in range(0, len(training)):
for i in range(0, len(training[dataset])):
mean = mean + torch.mean(training[dataset][i][0])
std = std + torch.std(training[dataset][i][0])
# print(mean)
print(f'{i} Mean: {mean / training_count}')
print(f'{i} Std: {std / training_count}')