预训练的卷积神经网络特征提取及应用

使用keras上的VGG16模型对ImageNet的训练结果进行特征提取，并在猫狗分类中应用，同时进行了数据增强。代码如下：

from keras import models
from keras import layers
from keras import optimizers
from keras.applications import VGG16
from keras.preprocessing.image import ImageDataGenerator
import matplotlib.pyplot as plt


conv_base = VGG16(weights='imagenet',
                  include_top=False,
                  input_shape=(150, 150, 3))
#建立模型
model = models.Sequential()
model.add(conv_base)
model.add(layers.Flatten())
model.add(layers.Dense(256, activation='relu'))
model.add(layers.Dense(1, activation='sigmoid'))
print(model.summary())
print(len(model.trainable_weights))
#冻结卷积基
conv_base.trainable = False
print(len(model.trainable_weights))
#猫狗图片集，训练集2000张，验证和测试集各1000张
train_dir = './datasets/train/'
validation_dir = './datasets/validation'
test_dir = './datasets/test'

#数据增强
train_datagen = ImageDataGenerator(

    rescale=1./255,
    rotation_range=40,
    width_shift_range=0.2,
    height_shift_range=0.2,
    shear_range=0.2,
    zoom_range=0.2,
    horizontal_flip=True,
    fill_mode='nearest'
)
test_datagen = ImageDataGenerator(rescale=1./255)

train_generator = train_datagen.flow_from_directory(
    train_dir,
    target_size=(150,150),
    batch_size=20,
    class_mode='binary'
)
validation_generator = test_datagen.flow_from_directory(
    validation_dir,
    target_size=(150,150),
    batch_size=20,
    class_mode='binary'
)

model.compile(optimizer=optimizers.RMSprop(lr=2e-5),
              loss='binary_crossentropy',
              metrics=['acc'])
history = model.fit_generator(
    train_generator,steps_per_epoch=100,
    epochs=30,
    validation_data=validation_generator,
    validation_steps=50
)
model.save('cat_and_dog_pre_train_gpu.h5')
acc = history.history['acc']
val_acc = history.history['val_acc']
loss = history.history['loss']
val_loss = history.history['val_loss']

epochs = range(1, len(acc)+1)
plt.plot(epochs, acc, 'bo', label='Traing acc')
plt.plot(epochs, val_acc, 'b', label='Validation acc')
plt.title('Training and validation accuracy')
plt.legend()
plt.figure()

plt.plot(epochs, loss, 'bo', label='Traing loss')
plt.plot(epochs, val_loss, 'b', label='Validation loss')
plt.title('Training and validation loss')
plt.legend()
plt.show()