Cifar100-Classifier/data_utils.py at master · saifelden/Cifar100-Classifier · GitHub

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
import numpy as np
import keras.datasets.mnist as mst

import keras.datasets.cifar100 as cf100
import tensorflow as tf
from tensorflow.examples.tutorials.mnist import input_data
class Data_utils:

	def load_mnist_data(self,source,path = ""):

		if source == "keras":
			"""
			load data from keras
			"""

			training,testing = mst.load_data()
			training_input,training_output = training
			testing_input,testing_output = testing
			training_input = training_input.reshape(-1,28,28,1)
			training_output = training_output.reshape(-1,1)
			testing_input = testing_input.reshape(-1,28,28,1)
			testing_output = testing_output.reshape(-1,1)


			validation_input = testing_input[5000:]
			testing_input = testing_input[0:5000]
			validation_output = testing_output[5000:]
			testing_output = testing_output[0:5000]


		elif source =="tensorflow":
			"""
			load data from tensorflow
			"""
			mnist = input_data.read_data_sets('MNIST_data', one_hot=True)
			training,validation,testing = mnist
			training_input = training.images
			training_output = training.labels
			testing_input = testing.images
			testing_output = testing.labels
			validation_input = validation.images
			validation_output = validation.labels

		elif source == "input file":
			"""
			used to talk the data from input file
			"""

		else:
			print("This data source is not supported")
			return

		return training_input,training_output,validation_input,validation_output,testing_input,testing_output

	def load_cifar100_data(self,source,path = ""):


		if source == "keras":
			"""
			load data from keras
			"""

			training,testing = cf100.load_data()
			training_input,training_output = training
			testing_input,testing_output = testing
			training_input = training_input.reshape(-1,32,32,3)
			training_output = training_output.reshape(-1,1)
			testing_input = testing_input.reshape(-1,32,32,3)
			testing_output = testing_output.reshape(-1,1)

			validation_input = testing_input[5000:]
			testing_input = testing_input[0:5000]

			validation_output = testing_output[5000:]
			testing_output = testing_output[0:5000]

		elif source == "input file":
			"""
			used to talk the data from input file
			"""

		else:
			print("This data source is not supported")
			return


		return training_input,training_output,validation_input,validation_output,testing_input,testing_output


	def apply_grey_scale(self,images):

		grayed_tf = tf.image.rgb_to_grayscale(images)
		sess = tf.Session()
		gray_images = sess.run(grayed_tf)

		gray_images.reshape(-1,images.shape[1],images.shape[2],1)
		reshaped = np.concatenate((gray_images,gray_images,gray_images),3)
		sess.close()
		return reshaped


	def apply_preprocessing(self,images,output,operations = ['gray','flip_vertical','flip_horizontal','transpose']):

		flipped_virtical_images = None
		all_images = images
		all_output = output
		if 'flip_vertical' in operations:
			flipped_virtical_images = np.flip(images,1)
			all_images = np.concatenate((all_images,flipped_virtical_images),0)
			all_output = np.concatenate((all_output,output),0)
		if 'flip_horizontal' in operations:
			flipped_horizontal_images = np.flip(images,2)
			all_images = np.concatenate((all_images,flipped_horizontal_images),0)
			all_output =  np.concatenate((all_output,output),0)
		if 'transpose' in operations:
			transposed_images = images.transpose(0,2,1,3)
			all_images = np.concatenate((all_images,transposed_images),0)
			all_output = np.concatenate((all_output,output),0)
		if 'gray' in operations:
			gray_images = self.apply_grey_scale(images)
			all_images = np.concatenate((all_images,gray_images),0)
			all_output = np.concatenate((all_output,output),0)

		return all_images,all_output