SortingImagesWithKeras/SortingImagesWithKeras.py at main · JiyoungMa/SortingImagesWithKeras · 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
137
import tensorflow.keras.preprocessing as kp
import numpy as np
from tensorflow.keras.models import Sequential, load_model
from tensorflow.keras.layers import Input, Conv2D, MaxPooling2D, Dense, Flatten, BatchNormalization, Dropout, GlobalAveragePooling2D
from sklearn.metrics import confusion_matrix, classification_report
import matplotlib.pyplot as plt
import tensorflow.keras.callbacks as callbacks

callback = callbacks.EarlyStopping(
    monitor = 'val_loss', min_delta = 0.001, patience = 3, mode = 'min'
)

train_generator = kp.image_dataset_from_directory(
 '/content/images',
 labels = "inferred",
 label_mode = 'categorical',
 class_names = ['food', 'interior', 'exterior'],
 color_mode="rgb",
 image_size=(300,300),
 batch_size=128,
 seed = 123,
 validation_split= 0.2,
 subset='training',
 interpolation = 'bilinear'
)

test_generator = kp.image_dataset_from_directory(
 '/content/images',
 labels="inferred",
 label_mode = 'categorical',
 color_mode="rgb",
 class_names = ['food', 'interior', 'exterior'],
 image_size=(300, 300),
 batch_size=128,
 seed = 123,
 validation_split=0.2,
 subset='validation',
 interpolation = 'bilinear')


batch_index = 0
test_batch = 0

model = Sequential([
        Input(shape=(300,300,3), name = 'input_layer'),
        BatchNormalization(),
        Conv2D(32, kernel_size=4,  strides= 1,activation='relu', name = 'conv_layer1'),
        MaxPooling2D(pool_size=2),
        BatchNormalization(),
        Conv2D(64, kernel_size=3, strides=1, activation='relu', name = 'conv_layer2'),
        MaxPooling2D(pool_size=2),
        BatchNormalization(),
        Conv2D(128, kernel_size=3, strides=1, activation='relu', name = 'conv_layer3'),
        MaxPooling2D(pool_size=2),
        BatchNormalization(),
        Conv2D(256, kernel_size=3, strides=1, activation='relu', name = 'conv_layer4'),
        MaxPooling2D(pool_size=2),
        BatchNormalization(),
        GlobalAveragePooling2D(),
        Dense(50, activation='relu', name = 'Dense_layer1'),
        Dense(100, activation='relu', name = 'Dense_layer2'),
        Dense(3, activation='softmax', name = 'output_layer')
    ])
model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])

model.summary()


history = model.fit(train_generator,batch_size=128, epochs = 10, validation_data=test_generator, validation_batch_size=128, callbacks = callback)
model.save('finalmodel')
def plot_loss_curve(history):
    import matplotlib.pyplot as plt

    plt.figure(figsize=(15,10))
    plt.plot(history['loss'])
    plt.plot(history['val_loss'])
    plt.title('model_loss')
    plt.ylabel('loss')
    plt.xlabel('epoch')
    plt.legend(['train', 'test'], loc = 'upper right')
    plt.show()

    plt.figure(figsize=(15, 10))
    plt.plot(history['accuracy'])
    plt.plot(history['val_accuracy'])
    plt.title('model_accuracy')
    plt.ylabel('accuracy')
    plt.xlabel('epoch')
    plt.legend(['train', 'test'], loc='upper right')
    plt.show()

plot_loss_curve(history.history)

#model = load_model('C:\\Users\\user\\Desktop\\3학년_2학기\\데이터\\model-201811250\\model-201811250')
test_generator = kp.image_dataset_from_directory(
 'C:\\Users\\user\\Desktop\\3학년_2학기\\데이터\\기말고사 대체과제\\images',
 labels="inferred",
 label_mode = 'categorical',
 color_mode="rgb",
 class_names = ['food', 'interior', 'exterior'],
 image_size=(300, 300),
 batch_size=128,
 seed = 123,
 validation_split=0.2,
 subset='validation',
 interpolation = 'bilinear')

predict = []
y = []
a = 1

for images, labels in test_generator:
  y_pred = model.predict(images,batch_size = 128)
  y_pred = np.argmax(y_pred, axis=1)
  labels = np.argmax(labels,axis=1)
  for i in range(len(y_pred)):
    #print(i)
    #plt.imshow(np.array(images[i]).astype('uint8'))
    #plt.show()
    print(labels[i] , " ", y_pred[i])
    y.append(labels[i])
    predict.append(y_pred[i])
    if labels[i] != y_pred[i]:
        print("sample %d is wrong!" % a)
        #plt.imshow(np.array(images[i]).astype('uint8'))
        #plt.show()
        #print("pause")
        with open("wrong_samples_fourteenth.txt", "a") as errfile:
            print("%d" % a, file=errfile)
    else:
        print("sample %d is correct" % a)
        #plt.imshow(np.array(images[i]).astype('uint8'))
        #plt.show()
        #print("pause")
    a += 1

print(classification_report(y, predict, target_names=['food','interior','exterior']))