Complex-Valued-SM/prepro.py at master · kwignb/Complex-Valued-SM · 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
import numpy as np
from PIL import Image
from sklearn.preprocessing import MinMaxScaler
from scipy.signal import hilbert, chirp


def unpickle(f):
    import _pickle as cPickle
    fo = open(f, 'rb')
    d = cPickle.load(fo, encoding='latin1')
    fo.close()
    return d


def load_dataset(data_dir):
    train_img = []
    train_label = []

    # load train data
    for i in range(1, 6):
        train_data = unpickle("%s/data_batch_%d" % (data_dir, i))
        train_img.extend(train_data["data"])
        train_label.extend(train_data["labels"])

    # load test data
    test_data = unpickle("%s/test_batch" % (data_dir))
    test_img = test_data["data"]
    test_label = test_data["labels"]

    # transform images to ndarray of float32 and labels to ndarray of int32
    train_img = np.array(train_img, dtype=np.float32)
    train_label = np.array(train_label, dtype=np.int32)
    test_img = np.array(test_img, dtype=np.float32)
    test_label = np.array(test_label, dtype=np.int32)

    return train_img, test_img, train_label, test_label


def load_dataset_ex(data_dir):
    train_img = []
    train_label = []

    # load train data (data_batch_1, 10000)
    train_data = unpickle("%s/data_batch_1" % (data_dir))
    train_img.extend(train_data["data"])
    train_label.extend(train_data["labels"])

    # load test data
    test_data = unpickle("%s/test_batch" % (data_dir))
    test_img = test_data["data"]
    test_label = test_data["labels"]

    # transform images to ndarray of float32 and labels to ndarray of int32
    train_img = np.array(train_img, dtype=np.float32)
    train_label = np.array(train_label, dtype=np.int32)
    test_img = np.array(test_img, dtype=np.float32)
    test_label = np.array(test_label, dtype=np.int32)

    return train_img, test_img, train_label, test_label


def monotone(train_img, test_img):
    train_data = []
    for i in range(len(train_img)):
        X_train_ = train_img[i].reshape((3, 32, 32)).transpose(1, 2, 0)
        # transform train color images to monochrome image
        X_train_ = np.array(Image.fromarray(np.uint8(X_train_)).convert("L")) / 255
        train_data.append(X_train_)

    test_data = []
    for i in range(len(test_img)):
        X_test_ = test_img[i].reshape((3, 32, 32)).transpose(1, 2, 0)
        # transform test color image to monochrome image
        X_test_ = np.array(Image.fromarray(np.uint8(X_test_)).convert("L")) / 255
        test_data.append(X_test_)

    return train_data, test_data


def hilbert_transform(train_img, test_img):
    train_data = []
    for i in range(len(train_img)):
        # transform train images to time series like
        train_signal = train_img[i].reshape(32 * 32, 1)
        train_signal = np.array(train_signal, dtype="float64")

        # normalize time series to the range [0,1]
        mms_train = MinMaxScaler()
        train_signal = mms_train.fit_transform(train_signal)
        # do Hilbert transform to the time series and transform time series to image
        train_signal = hilbert(train_signal, axis=0).reshape(32, 32)
        train_data.append(train_signal)

    test_data = []
    for i in range(len(test_img)):
        # transform test images to time series like
        test_signal = test_img[i].reshape(32 * 32, 1)
        test_signal = np.array(test_signal, dtype="float64")

        # normalize time series to the range [0,1]
        mms_test = MinMaxScaler()
        test_signal = mms_test.fit_transform(test_signal)
        # do Hilbert transform to the time series and transform time series to image
        test_signal = hilbert(test_signal, axis=0).reshape(32, 32)
        test_data.append(test_signal)

    return train_data, test_data


def monotone_i(train_img, test_img):
    train_data = []
    for i in range(len(train_img)):
        X_train_ = train_img[i].reshape((3, 32, 32)).transpose(1, 2, 0)
        # transform train color images to monochrome image
        X_train_ = np.array(Image.fromarray(np.uint8(X_train_)).convert("L")) / 255 + 0.01j
        train_data.append(X_train_)

    test_data = []
    for i in range(len(test_img)):
        X_test_ = test_img[i].reshape((3, 32, 32)).transpose(1, 2, 0)
        # transform test color image to monochrome image
        X_test_ = np.array(Image.fromarray(np.uint8(X_test_)).convert("L")) / 255 + 0.01j
        test_data.append(X_test_)

    return train_data, test_data