使用预训练网络
- 在小数据集上使用深度学习,除了自己从头开始训练,另一种高效的方法是使用别人在大型数据集上预训练好的模型。
from keras.applications import VGG16
conv_base = VGG16(weights='imagenet',
include_top=False, #模型中是否包含全连接的分类器
input_shape=(150,150,3))
Using TensorFlow backend.
conv_base.summary()
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_1 (InputLayer) (None, 150, 150, 3) 0
_________________________________________________________________
block1_conv1 (Conv2D) (None, 150, 150, 64) 1792
_________________________________________________________________
block1_conv2 (Conv2D) (None, 150, 150, 64) 36928
_________________________________________________________________
block1_pool (MaxPooling2D) (None, 75, 75, 64) 0
_________________________________________________________________
block2_conv1 (Conv2D) (None, 75, 75, 128) 73856
_________________________________________________________________
block2_conv2 (Conv2D) (None, 75, 75, 128) 147584
_________________________________________________________________
block2_pool (MaxPooling2D) (None, 37, 37, 128) 0
_________________________________________________________________
block3_conv1 (Conv2D) (None, 37, 37, 256) 295168
_________________________________________________________________
block3_conv2 (Conv2D) (None, 37, 37, 256) 590080
_________________________________________________________________
block3_conv3 (Conv2D) (None, 37, 37, 256) 590080
_________________________________________________________________
block3_pool (MaxPooling2D) (None, 18, 18, 256) 0
_________________________________________________________________
block4_conv1 (Conv2D) (None, 18, 18, 512) 1180160
_________________________________________________________________
block4_conv2 (Conv2D) (None, 18, 18, 512) 2359808
_________________________________________________________________
block4_conv3 (Conv2D) (None, 18, 18, 512) 2359808
_________________________________________________________________
block4_pool (MaxPooling2D) (None, 9, 9, 512) 0
_________________________________________________________________
block5_conv1 (Conv2D) (None, 9, 9, 512) 2359808
_________________________________________________________________
block5_conv2 (Conv2D) (None, 9, 9, 512) 2359808
_________________________________________________________________
block5_conv3 (Conv2D) (None, 9, 9, 512) 2359808
_________________________________________________________________
block5_pool (MaxPooling2D) (None, 4, 4, 512) 0
=================================================================
Total params: 14,714,688
Trainable params: 14,714,688
Non-trainable params: 0
_________________________________________________________________
- 可以看出最后输出的特征图的大小是(4,4,512),下面需要在这个特征上添加一个密集连接的分类器,有两种方法:
- A:在自己的数据集上运行卷积基,将输出的结果保存在本地硬盘的numpy数组中,然后用这些数据作为输入,输入到一个独立的全连接分类器中。
- 特点:速度快,计算代价低,但是不可以使用数据增强技术
- B:在conv_base模型的基础上直接添加Dense层,扩展已有的模型,并在输入数据上端到端的运行整个模型。
- 特点:可以使用数据增强技术,但是计算的代价较高。
- A:在自己的数据集上运行卷积基,将输出的结果保存在本地硬盘的numpy数组中,然后用这些数据作为输入,输入到一个独立的全连接分类器中。
1、A方案
import os
import numpy as np
from keras.preprocessing.image import ImageDataGenerator
base_dir = r"F:\Data_Set"
train_dir = os.path.join(base_dir, 'train_dir')
validation_dir = os.path.join(base_dir, 'validation_dir')
test_dir = os.path.join(base_dir, 'test_dir')
datagen = ImageDataGenerator(rescale=1./255)
batch_size = 20
def extract_features(directory, sample_count):
features = np.zeros(shape=(sample_count, 4,4,512))
labels = np.zeros(shape=(sample_count))
generator = datagen.flow_from_directory(directory,
target_size=(150,150),
batch_size=batch_size,
class_mode="binary")
i = 0
for input_batch, labels_batch in generator:
features_batch = conv_base.predict(input_batch)
features[i * batch_size:(i + 1) * batch_size] = features_batch
labels[i * batch_size:(i+1) * batch_size] = labels_batch
i += 1
if i* batch_size >= sample_count:
break
return features, labels
train_features, train_labels = extract_features(train_dir, 2000)
validation_features, validation_labels = extract_features(validation_dir, 1000)
test_features, test_labels = extract_features(test_dir, 1000)
Found 2000 images belonging to 2 classes.
- 上面的代码可能是因为电脑内存的原因导致运行不成功
- 现在假设上面的代码已经将输入的特征提取出来了,其大小为(samples, 4,4,512)
- 我们将其flatten,训练一个二分类的分类器
train_features = np.reshape(train_features, (2000, 4,4,512))
validation_features = np.reshape(validation_features,(1000,4,4,512))
test_features = np.reshape(test_features, (1000,4,4512))
- 上面的代码相当于为我们准备好了输入的特征数据
- 搭建分类模型
from keras import models
from keras import layers
from keras import optimizers
model = models.Sequential()
model.add(layers.Dense(256, activation='relu', input_dim=4*4*512))
model.add(layers.Dropout(0.5))
model.add(layers.Dense(1,activation='sigmoid'))
model.compile(optimizer=optimizers.RMSprop(lr=2e-5),
loss='binary_crossentropy',
metrics=['acc'])
history = model.fit(train_features, train_labels,
epochs=30,
batch_size=20,
validation_data=(validation_features, validation_labels))
2、B方案
from keras import models
from keras import layers
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'))
model.summary()
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
vgg16 (Model) (None, 4, 4, 512) 14714688
_________________________________________________________________
flatten_1 (Flatten) (None, 8192) 0
_________________________________________________________________
dense_5 (Dense) (None, 256) 2097408
_________________________________________________________________
dense_6 (Dense) (None, 1) 257
=================================================================
Total params: 16,812,353
Trainable params: 16,812,353
Non-trainable params: 0
_________________________________________________________________
- 可以看到VGG16的参数非常多,在编译和训练模型之前,一定要冻结卷积基。
- 冻结: 是指一个层或者多个层在训练的过程中保持其权重不变。
print("在冻结之前可训练的权重张量的个数: ",len(model.trainable_weights))
conv_base.trainable = False
print("冻结之后可训练的权重张量的个数: ", len(model.trainable_weights))
在冻结之前可训练的权重张量的个数: 30
冻结之后可训练的权重张量的个数: 4
- 可以看到在冻结之后,只有添加的两个全连接层的权重才会被训练。每层两个权重张量。
- 接下来利用冻结的卷积基端到端的训练模型
from keras.preprocessing.image import ImageDataGenerator
from keras import optimizers
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")
tset_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(loss='binary_crossentropy',
optimizer=optimizers.RMSprop(lr=2e-5),
metrics=['acc'])
history = model.fit_generator(train_generator,
step_per_epoch=100,
epochs=30,
validation_data=validation_generator,
validation_steps=50)
卷积神经网络的可视化
- 可视化中间激活: 对于指定的输入,展示网络中各个卷积层和池化层的输出特征图(层的输出通常被称为该层的激活,即激活函数的输出)
from keras.models import load_model
model = load_model("model_weights/cats_and_dogs_samll_2.h5")
model.summary()
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
conv2d_5 (Conv2D) (None, 148, 148, 32) 896
_________________________________________________________________
max_pooling2d_5 (MaxPooling2 (None, 74, 74, 32) 0
_________________________________________________________________
conv2d_6 (Conv2D) (None, 72, 72, 64) 18496
_________________________________________________________________
max_pooling2d_6 (MaxPooling2 (None, 36, 36, 64) 0
_________________________________________________________________
conv2d_7 (Conv2D) (None, 34, 34, 128) 73856
_________________________________________________________________
max_pooling2d_7 (MaxPooling2 (None, 17, 17, 128) 0
_________________________________________________________________
conv2d_8 (Conv2D) (None, 15, 15, 128) 147584
_________________________________________________________________
max_pooling2d_8 (MaxPooling2 (None, 7, 7, 128) 0
_________________________________________________________________
flatten_2 (Flatten) (None, 6272) 0
_________________________________________________________________
dropout_1 (Dropout) (None, 6272) 0
_________________________________________________________________
dense_3 (Dense) (None, 512) 3211776
_________________________________________________________________
dense_4 (Dense) (None, 1) 513
=================================================================
Total params: 3,453,121
Trainable params: 3,453,121
Non-trainable params: 0
_________________________________________________________________
# 使用一张图像
img_path = r"F:\Data_Set\test_dir\test_cats_dir\cat.1502.jpg"
from keras.preprocessing import image
import numpy as np
img = image.load_img(img_path, target_size=(150,150))
img_tensor = image.img_to_array(img)
img_tensor = np.expand_dims(img_tensor, axis=0)
img_tensor /= 255.
print(img_tensor.shape)
(1, 150, 150, 3)
import matplotlib.pyplot as plt
plt.imshow(img_tensor[0])
plt.show()
from keras import models
layer_outputs = [layer.output for layer in model.layers[:8]]# 提取前8层的输出
activation_model = models.Model(inputs=model.input, outputs=layer_outputs)# 创建一个模型,给定一个输入,可以返回各层的输出
activations = activation_model.predict(img_tensor) # 返回8个numpy数组组成的list,每层的输出对应一个数组
first_layer_activation = activations[0]
print(first_layer_activation.shape)
(1, 148, 148, 32)
plt.matshow(first_layer_activation[0,:,:,4], cmap='viridis')
<matplotlib.image.AxesImage at 0x1ef74c397f0>
plt.matshow(first_layer_activation[0,:,:,10], cmap='viridis')
<matplotlib.image.AxesImage at 0x1ef88121588>
- 接下来展示所有的特征图
#获取层的名称
layer_names = []
for layer in model.layers[:8]:
layer_names.append(layer.name)
images_per_row = 16
for layer_name, layer_activation in zip(layer_names, activations):
n_features = layer_activation.shape[-1]
size = layer_activation.shape[1]
n_cols = n_features // images_per_row
display_grid = np.zeros((size * n_cols, images_per_row * size ))
for col in range(n_cols):
for row in range(images_per_row):
channel_image = layer_activation[0, :, :, col * images_per_row + row]
channel_image -= channel_image.mean()
channel_image /= channel_image.std()
channel_image *= 64
channel_image += 128
channel_image = np.clip(channel_image, 0,255).astype('uint8')
display_grid[col * size : (col + 1) * size,
row * size : (row + 1) * size] = channel_image
scale = 1./size
plt.figure(figsize=(scale * display_grid.shape[1],
scale * display_grid.shape[0]))
plt.title(layer_name)
plt.grid(False)
plt.imshow(display_grid,aspect='auto',cmap='viridis')
D:\Anaconda3\envs\GPU\lib\site-packages\matplotlib\pyplot.py:513: RuntimeWarning: More than 20 figures have been opened. Figures created through the pyplot interface (`matplotlib.pyplot.figure`) are retained until explicitly closed and may consume too much memory. (To control this warning, see the rcParam `figure.max_open_warning`).
max_open_warning, RuntimeWarning)
layer_names
['conv2d_5',
'max_pooling2d_5',
'conv2d_6',
'max_pooling2d_6',
'conv2d_7',
'max_pooling2d_7',
'conv2d_8',
'max_pooling2d_8']
显示激活的热力图
from keras.applications.vgg16 import VGG16
from keras.preprocessing import image
from keras.applications.vgg16 import preprocess_input,decode_predictions
import numpy as np
img_path = r"F:\Data_Set\elephant.jpg"
img = image.load_img(img_path, target_size=(224,224))
x = image.img_to_array(img)
x = np.expand_dims(x, axis=0)
x = preprocess_input(x)
print(x.shape)
Using TensorFlow backend.
(1, 224, 224, 3)
model = VGG16(weights='imagenet')
preds = model.predict(x)
print(decode_predictions(preds, top=3)[0])
[('n02504458', 'African_elephant', 0.8086415), ('n01871265', 'tusker', 0.17776239), ('n02504013', 'Indian_elephant', 0.013450568)]
np.argmax(preds[0])
386
- 下面展示图像中的哪些部分最像非洲象,使用Grad_CAM
import keras.backend as K
import matplotlib.pyplot as plt
african_elephant_output = model.output[:, 386]
last_conv_layer = model.get_layer('block5_conv3')
grads = K.gradients(african_elephant_output, last_conv_layer.output)[0]
pooled_grads = K.mean(grads, axis=(0,1,2))
iterate = K.function([model.input],
[pooled_grads, last_conv_layer.output[0]])
pooled_grads_value, conv_layer_output_value = iterate([x])
for i in range(512):
conv_layer_output_value[:,:,i] *= pooled_grads_value[i]
heatmap = np.mean(conv_layer_output_value, axis=-1)
heatmap = np.maximum(heatmap, 0)
heatmap /= np.max(heatmap)
plt.matshow(heatmap)
<matplotlib.image.AxesImage at 0x11dace6beb8>
- 最后使用opencv生成一张图像将原始图像添加在刚得到的热力图上
import cv2
img = cv2.imread(img_path)
print(img.shape)
print(heatmap.shape)
heatmap = np.uint8(255 * heatmap)
heatmap = cv2.resize(heatmap, (img.shape[1], img.shape[0]))
heatmap = cv2.applyColorMap(heatmap, cv2.COLORMAP_JET)
superimposed_img = heatmap * 0.4 + img
cv2.imwrite(r"F:\Data_Set\elephant_heatmap.jpg", superimposed_img)
(443, 666, 3)
(443, 666, 3)
True
补充: 完整的VGG16模型结构
model2=VGG16()
Downloading data from https://github.com/fchollet/deep-learning-models/releases/download/v0.1/vgg16_weights_tf_dim_ordering_tf_kernels.h5
553467904/553467096 [==============================] - 522s 1us/step
model2.summary()
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_2 (InputLayer) (None, 224, 224, 3) 0
_________________________________________________________________
block1_conv1 (Conv2D) (None, 224, 224, 64) 1792
_________________________________________________________________
block1_conv2 (Conv2D) (None, 224, 224, 64) 36928
_________________________________________________________________
block1_pool (MaxPooling2D) (None, 112, 112, 64) 0
_________________________________________________________________
block2_conv1 (Conv2D) (None, 112, 112, 128) 73856
_________________________________________________________________
block2_conv2 (Conv2D) (None, 112, 112, 128) 147584
_________________________________________________________________
block2_pool (MaxPooling2D) (None, 56, 56, 128) 0
_________________________________________________________________
block3_conv1 (Conv2D) (None, 56, 56, 256) 295168
_________________________________________________________________
block3_conv2 (Conv2D) (None, 56, 56, 256) 590080
_________________________________________________________________
block3_conv3 (Conv2D) (None, 56, 56, 256) 590080
_________________________________________________________________
block3_pool (MaxPooling2D) (None, 28, 28, 256) 0
_________________________________________________________________
block4_conv1 (Conv2D) (None, 28, 28, 512) 1180160
_________________________________________________________________
block4_conv2 (Conv2D) (None, 28, 28, 512) 2359808
_________________________________________________________________
block4_conv3 (Conv2D) (None, 28, 28, 512) 2359808
_________________________________________________________________
block4_pool (MaxPooling2D) (None, 14, 14, 512) 0
_________________________________________________________________
block5_conv1 (Conv2D) (None, 14, 14, 512) 2359808
_________________________________________________________________
block5_conv2 (Conv2D) (None, 14, 14, 512) 2359808
_________________________________________________________________
block5_conv3 (Conv2D) (None, 14, 14, 512) 2359808
_________________________________________________________________
block5_pool (MaxPooling2D) (None, 7, 7, 512) 0
_________________________________________________________________
flatten (Flatten) (None, 25088) 0
_________________________________________________________________
fc1 (Dense) (None, 4096) 102764544
_________________________________________________________________
fc2 (Dense) (None, 4096) 16781312
_________________________________________________________________
predictions (Dense) (None, 1000) 4097000
=================================================================
Total params: 138,357,544
Trainable params: 138,357,544
Non-trainable params: 0
_________________________________________________________________