-
Notifications
You must be signed in to change notification settings - Fork 0
/
DataSet.py
172 lines (146 loc) · 6.99 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
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
"""
Construct a NeuralNetwork class to include operations
related to various datasets and corresponding models.
Author: Min Wu
Email: min.wu@cs.ox.ac.uk
"""
import keras
from keras.datasets import mnist, cifar10, fashion_mnist
from skimage import io, color, exposure, transform
import pandas as pd
import numpy as np
import h5py
import os
import glob
# Define a Neural Network class.
class DataSet:
# Specify which dataset at initialisation.
def __init__(self, data_set, trainOrTest):
self.data_set = data_set
# for a mnist model.
if self.data_set == 'mnist':
num_classes = 10
(x_train, y_train), (x_test, y_test) = mnist.load_data()
img_rows, img_cols, img_chls = 28, 28, 1
if trainOrTest == "training":
x = x_train.reshape(x_train.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_train, num_classes)
else:
x = x_test.reshape(x_test.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_test, num_classes)
y= 2*y-1
x = x.astype('float64')
x /= 255
elif self.data_set == 'fashion':
num_classes = 10
(x_train, y_train), (x_test, y_test) = fashion_mnist.load_data()
img_rows, img_cols, img_chls = 28, 28, 1
if trainOrTest == "training":
x = x_train.reshape(x_train.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_train, num_classes)
else:
x = x_test.reshape(x_test.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_test, num_classes)
y= 2*y-1
x = x.astype('float64')
x /= 255
# for a cifar10 model.
elif self.data_set == 'cifar10':
num_classes = 10
(x_train, y_train), (x_test, y_test) = cifar10.load_data()
img_rows, img_cols, img_chls = 32, 32, 3
if trainOrTest == "training":
x = x_train.reshape(x_train.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_train, num_classes)
else:
x = x_test.reshape(x_test.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_test, num_classes)
y = 2*y-1
x = x.astype('float64')
x /= 255
# for a gtsrb model.
elif self.data_set == 'gtsrb':
num_classes = 43
img_rows, img_cols, img_chls = 48, 48, 3
if trainOrTest == "training":
directory = 'models/GTSRB/Final_Training/'
try:
with h5py.File(directory + 'gtsrb_training.h5') as hf:
x_train, y_train = hf['imgs'][:], hf['labels'][:]
x = x_train.reshape(x_train.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_train, num_classes)
except (IOError, OSError, KeyError):
imgs = []
labels = []
all_img_paths = glob.glob(os.path.join(directory + 'Images/', '*/*.ppm'))
np.random.shuffle(all_img_paths)
for img_path in all_img_paths:
try:
img = self.preprocess_img(io.imread(img_path), img_rows, img_cols)
label = self.get_class(img_path)
imgs.append(img)
labels.append(label)
if len(imgs) % 1000 == 0: print("Processed {}/{}".format(len(imgs), len(all_img_paths)))
except (IOError, OSError):
print('missed', img_path)
pass
x_train = np.array(imgs, dtype='float64')
y_train = np.array(labels, dtype='uint8')
with h5py.File(directory + 'gtsrb_training.h5', 'w') as hf:
hf.create_dataset('imgs', data=x_train)
hf.create_dataset('labels', data=y_train)
x = x_train.reshape(x_train.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_train, num_classes)
else:
directory = 'models/GTSRB/Final_Test/'
try:
with h5py.File(directory + 'gtsrb_test.h5') as hf:
x_test, y_test = hf['imgs'][:], hf['labels'][:]
x = x_test.reshape(x_test.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_test, num_classes)
except (IOError, OSError, KeyError):
test = pd.read_csv(directory + 'GT-final_test.csv', sep=';')
x_test = []
y_test = []
for file_name, class_id in zip(list(test['Filename']), list(test['ClassId'])):
img_path = os.path.join(directory + 'Images/', file_name)
x_test.append(self.preprocess_img(io.imread(img_path), img_rows, img_cols))
y_test.append(class_id)
x_test = np.array(x_test, dtype='float64')
y_test = np.array(y_test, dtype='uint8')
with h5py.File(directory + 'gtsrb_test.h5', 'w') as hf:
hf.create_dataset('imgs', data=x_test)
hf.create_dataset('labels', data=y_test)
x = x_test.reshape(x_test.shape[0], img_rows, img_cols, img_chls)
y = keras.utils.np_utils.to_categorical(y_test, num_classes)
else:
print("Unsupported dataset %s. Try 'mnist' or 'cifar10'." % data_set)
exit()
self.x = x
self.y = y
# get dataset
def get_dataset(self):
return self.x, self.y
def get_input(self, index):
return self.x[index]
def get_True_Label(self, index):
arrayLabels=self.y[index]
indexs=np.where( arrayLabels > 0 )[0]
return indexs[0]
def preprocess_img(self, img, img_rows, img_cols):
# Histogram normalization in y
hsv = color.rgb2hsv(img)
hsv[:, :, 2] = exposure.equalize_hist(hsv[:, :, 2])
img = color.hsv2rgb(hsv)
# central scrop
min_side = min(img.shape[:-1])
centre = img.shape[0] // 2, img.shape[1] // 2
img = img[centre[0] - min_side // 2:centre[0] + min_side // 2,
centre[1] - min_side // 2:centre[1] + min_side // 2, :]
# rescale to standard size
img = transform.resize(img, (img_rows, img_cols))
# roll color axis to axis 0
# img = np.rollaxis(img, -1)
return img
def get_class(self, img_path):
return int(img_path.split('/')[-2])