def densenet161_model(img_rows,
img_cols,
color_type=1,
nb_dense_block=4,
growth_rate=48,
nb_filter=96,
reduction=0.5,
dropout_rate=0.0,
weight_decay=1e-4,
num_classes=None):
'''
DenseNet 161 Model for Keras
Model Schema is based on
https://github.com/flyyufelix/DenseNet-Keras
ImageNet Pretrained Weights
Theano: https://drive.google.com/open?id=0Byy2AcGyEVxfVnlCMlBGTDR3RGs
TensorFlow: https://drive.google.com/open?id=0Byy2AcGyEVxfUDZwVjU2cFNidTA
# Arguments
nb_dense_block: number of dense blocks to add to end
growth_rate: number of filters to add per dense block
nb_filter: initial number of filters
reduction: reduction factor of transition blocks.
dropout_rate: dropout rate
weight_decay: weight decay factor
classes: optional number of classes to classify images
weights_path: path to pre-trained weights
# Returns
A Keras model instance.
'''
eps = 1.1e-5
# compute compression factor
compression = 1.0 - reduction
# Handle Dimension Ordering for different backends
global concat_axis
if K.image_dim_ordering() == 'tf':
concat_axis = 3
img_input = Input(shape=(224, 224, 3), name='data')
else:
concat_axis = 1
img_input = Input(shape=(3, 224, 224), name='data')
# From architecture for ImageNet (Table 1 in the paper)
nb_filter = 96
nb_layers = [6, 12, 36, 24] # For DenseNet-161
# Initial convolution
x = ZeroPadding2D((3, 3), name='conv1_zeropadding')(img_input)
x = Convolution2D(nb_filter,
7,
7,
subsample=(2, 2),
name='conv1',
bias=False)(x)
x = BatchNormalization(epsilon=eps, axis=concat_axis, name='conv1_bn')(x)
x = Scale(axis=concat_axis, name='conv1_scale')(x)
x = Activation('relu', name='relu1')(x)
x = ZeroPadding2D((1, 1), name='pool1_zeropadding')(x)
x = MaxPooling2D((3, 3), strides=(2, 2), name='pool1')(x)
# Add dense blocks
for block_idx in range(nb_dense_block - 1):
stage = block_idx + 2
x, nb_filter = dense_block(x,
stage,
nb_layers[block_idx],
nb_filter,
growth_rate,
dropout_rate=dropout_rate,
weight_decay=weight_decay)
# Add transition_block
x = transition_block(x,
stage,
nb_filter,
compression=compression,
dropout_rate=dropout_rate,
weight_decay=weight_decay)
nb_filter = int(nb_filter * compression)
final_stage = stage + 1
x, nb_filter = dense_block(x,
final_stage,
nb_layers[-1],
nb_filter,
growth_rate,
dropout_rate=dropout_rate,
weight_decay=weight_decay)
x = BatchNormalization(epsilon=eps,
axis=concat_axis,
name='conv' + str(final_stage) + '_blk_bn')(x)
x = Scale(axis=concat_axis,
name='conv' + str(final_stage) + '_blk_scale')(x)
x = Activation('relu', name='relu' + str(final_stage) + '_blk')(x)
x_fc = GlobalAveragePooling2D(name='pool' + str(final_stage))(x)
x_fc = Dense(1000, name='fc6')(x_fc)
x_fc = Activation('softmax', name='prob')(x_fc)
model = Model(img_input, x_fc, name='densenet')
if K.image_dim_ordering() == 'th':
# Use pre-trained weights for Theano backend
weights_path = 'imagenet_models/densenet161_weights_th.h5'
else:
# Use pre-trained weights for Tensorflow backend
weights_path = 'imagenet_models/densenet161_weights_tf.h5'
# model.load_weights(weights_path, by_name=True)
# Truncate and replace softmax layer for transfer learning
# Cannot use model.layers.pop() since model is not of Sequential() type
# The method below works since pre-trained weights are stored in layers but not in the model
x_newfc = GlobalAveragePooling2D(name='pool' + str(final_stage))(x)
x_newfc = Dense(num_classes, name='fc6')(x_newfc)
x_newfc = Activation('softmax', name='prob')(x_newfc)
model = Model(img_input, x_newfc)
# Learning rate is changed to 0.001
sgd = SGD(lr=1e-3, decay=1e-6, momentum=0.9, nesterov=True)
model.compile(optimizer=sgd,
loss='categorical_crossentropy',
metrics=['accuracy'])
return model
def cnn_model02(img_rows, img_cols, color_type=1, num_classes=None):
eps = 1.1e-5
# 处理尺寸不同的后端
global bn_axis
# ## th : if image_dim_ordering = channels_first”数据组织为（3,128,128,128），
# ## tf : ...=“channels_last”数据组织为（128,128,128,3）
if K.image_data_format() == 'channels_last':
bn_axis = 3
img_input = Input(shape=(img_rows, img_cols, color_type), name='data')
else:
bn_axis = 1
img_input = Input(shape=(color_type, img_rows, img_cols), name='data')
x = ZeroPadding2D((3, 3), name='conv1_zeropadding')(img_input)
x = Conv2D(64, (7, 7), strides=(2, 2), name='conv1', use_bias=False)(x)
# x = GroupNormalization()(x)
x = BatchNormalization(epsilon=eps, axis=bn_axis, name='bn_conv1')(x)
x = Scale(axis=bn_axis, name='scale_conv1')(x)
x = Activation('relu', name='conv1_relu')(x)
x = MaxPooling2D((3, 3), strides=(2, 2), name='pool1')(x)
x = conv_block(x, 3, [64, 64, 256], stage=2, block='a', strides=(1, 1))
x = identity_block(x, 3, [64, 64, 256], stage=2, block='b')
x = identity_block(x, 3, [64, 64, 256], stage=2, block='c')
x = conv_block_D(x, 3, [128, 128, 512], stage=3, block='a')
for i in range(1, 8):
x = identity_block(x, 3, [128, 128, 512], stage=3, block='b' + str(i))
x = conv_block_D(x, 3, [256, 256, 1024], stage=4, block='a')
for i in range(1, 36):
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='b' + str(i))
x = conv_block_D(x, 3, [512, 512, 2048], stage=5, block='a')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='b')
x = identity_block_tanh(x, 3, [512, 512, 2048], stage=5, block='c')
x_fc = AveragePooling2D((7, 7), name='avg_pool')(x)
x_fc = Flatten()(x_fc)
x_fc = Dropout(0.1)(x_fc)
# Dense(units:输出维度，activation=激活函数)全连接层（对上一层的神经元进行全部连接，实现特征的非线性组合）
x_fc = Dense(1000, activation='softmax', name='fc1000')(x_fc)
model = Model(img_input, x_fc)
# net = get_model('ResNet50_v1d', pretrained='117a384e')
# you may modify it to switch to another model. The name is case-insensitive
# model_name = 'ResNet152_v1d'
# # download and load the pre-trained model
# net = gluoncv.model_zoo.get_model(model_name, pretrained='cddbc86f')
# net_params = net.collect_params()
if K.image_data_format() == 'channels_first':
# 使用预先训练过的权重进行Theano后端
# weights_path = 'ResNet152_v1d.h5'
weights_path = 'models/resnet152_weights_tf.h5'
else:
# 在Tensorflow后端使用预先训练的权重
# weights_path = 'ResNet152_v1d.h5'
weights_path = 'models/resnet152_weights_tf.h5'
model.load_weights(weights_path, by_name=True)
# 截断并替换softmax层以进行传输学习
# 不能使用model.layers.pop（），因为model不是Sequential（）类型
# 下面的方法有效，因为预训练的权重存储在图层中但不存储在模型中
x_newfc = AveragePooling2D((7, 7), name='avg_pool')(x)
x_newfc = Flatten()(x_newfc)
x_newfc = Dropout(0.1)(x_newfc)
x_newfc = Dense(num_classes, activation=None, name='fc8')(x_newfc)
model = Model(img_input, x_newfc)
# 学习率改为0.001
sgd = SGD(lr=1e-3, decay=1e-6, momentum=0.9, nesterov=True)
model.compile(optimizer=sgd,
loss=cl.loss,
metrics=[cl.categorical_accuracy])
return model
def resnet101_model(net_settings,
img_rows,
img_cols,
color_type=1,
num_classes=None):
"""
Resnet 101 Model for Keras
Model Schema and layer naming follow that of the original Caffe implementation
https://github.com/KaimingHe/deep-residual-networks
ImageNet Pretrained Weights
Theano: https://drive.google.com/file/d/0Byy2AcGyEVxfdUV1MHJhelpnSG8/view?usp=sharing
TensorFlow: https://drive.google.com/file/d/0Byy2AcGyEVxfTmRRVmpGWDczaXM/view?usp=sharing
Parameters:
net_settings : dictionary with configuration settings
img_rows, img_cols - resolution of inputs
channel - 1 for grayscale, 3 for color
num_classes - number of class labels for our classification task
"""
eps = 1.1e-5
# Handle Dimension Ordering for different backends
global bn_axis
if K.image_dim_ordering() == 'tf':
bn_axis = 3
img_input = Input(shape=(img_rows, img_cols, color_type), name='data')
else:
bn_axis = 1
img_input = Input(shape=(color_type, img_rows, img_cols), name='data')
x = ZeroPadding2D((3, 3), name='conv1_zeropadding')(img_input)
x = Convolution2D(64, 7, 7, subsample=(2, 2), name='conv1', bias=False)(x)
x = BatchNormalization(epsilon=eps, axis=bn_axis, name='bn_conv1')(x)
x = Scale(axis=bn_axis, name='scale_conv1')(x)
x = Activation('relu', name='conv1_relu')(x)
x = MaxPooling2D((3, 3), strides=(2, 2), name='pool1')(x)
x = conv_block(x, 3, [64, 64, 256], stage=2, block='a', strides=(1, 1))
x = identity_block(x, 3, [64, 64, 256], stage=2, block='b')
x = identity_block(x, 3, [64, 64, 256], stage=2, block='c')
x = conv_block(x, 3, [128, 128, 512], stage=3, block='a')
for i in range(1, 4):
x = identity_block(x, 3, [128, 128, 512], stage=3, block='b' + str(i))
x = conv_block(x, 3, [256, 256, 1024], stage=4, block='a')
for i in range(1, 23):
x = identity_block(x, 3, [256, 256, 1024], stage=4, block='b' + str(i))
x = conv_block(x, 3, [512, 512, 2048], stage=5, block='a')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='b')
x = identity_block(x, 3, [512, 512, 2048], stage=5, block='c')
x_fc = AveragePooling2D((7, 7), name='avg_pool')(x)
x_fc = Flatten()(x_fc)
x_fc = Dense(1000, activation='softmax', name='fc1000')(x_fc)
model = Model(img_input, x_fc)
if K.image_dim_ordering() == 'th':
# Use pre-trained weights for Theano backend
weights_path = 'imagenet_models/resnet101_weights_th.h5'
else:
# Use pre-trained weights for Tensorflow backend
weights_path = 'imagenet_models/resnet101_weights_tf.h5'
model.load_weights(weights_path, by_name=True)
# Truncate and replace softmax layer for transfer learning
# Cannot use model.layers.pop() since model is not of Sequential() type
# The method below works since pre-trained weights are stored in layers but not in the model
x_newfc = AveragePooling2D((7, 7), name='avg_pool')(x)
x_newfc = Flatten()(x_newfc)
x_newfc = Dense(num_classes,
activation=net_settings['activation'],
name='fc8',
kernel_regularizer=keras.regularizers.l2(0.01))(x_newfc)
model = Model(img_input, x_newfc)
# Learning rate is changed to 0.001
#sgd = SGD(lr=1e-4, decay=1e-3, momentum=0.9, nesterov=True)
#model.compile(optimizer=sgd, loss='binary_crossentropy', metrics=['accuracy'])
return model
def conv_block(input_tensor,
kernel_size,
filters,
stage,
block,
strides=(2, 2)):
'''conv_block is the block that has a conv layer at shortcut
# Arguments
input_tensor: input tensor
kernel_size: defualt 3, the kernel size of middle conv layer at main path
filters: list of integers, the nb_filters of 3 conv layer at main path
stage: integer, current stage label, used for generating layer names
block: 'a','b'..., current block label, used for generating layer names
Note that from stage 3, the first conv layer at main path is with subsample=(2,2)
And the shortcut should have subsample=(2,2) as well
'''
eps = 1.1e-5
nb_filter1, nb_filter2, nb_filter3 = filters
conv_name_base = 'res' + str(stage) + block + '_branch'
bn_name_base = 'bn' + str(stage) + block + '_branch'
scale_name_base = 'scale' + str(stage) + block + '_branch'
x = Convolution2D(nb_filter1,
1,
1,
subsample=strides,
name=conv_name_base + '2a',
bias=False)(input_tensor)
x = BatchNormalization(epsilon=eps, axis=bn_axis,
name=bn_name_base + '2a')(x)
x = Scale(axis=bn_axis, name=scale_name_base + '2a')(x)
x = Activation('relu', name=conv_name_base + '2a_relu')(x)
x = ZeroPadding2D((1, 1), name=conv_name_base + '2b_zeropadding')(x)
x = Convolution2D(nb_filter2,
kernel_size,
kernel_size,
name=conv_name_base + '2b',
bias=False)(x)
x = BatchNormalization(epsilon=eps, axis=bn_axis,
name=bn_name_base + '2b')(x)
x = Scale(axis=bn_axis, name=scale_name_base + '2b')(x)
x = Activation('relu', name=conv_name_base + '2b_relu')(x)
x = Convolution2D(nb_filter3, 1, 1, name=conv_name_base + '2c',
bias=False)(x)
x = BatchNormalization(epsilon=eps, axis=bn_axis,
name=bn_name_base + '2c')(x)
x = Scale(axis=bn_axis, name=scale_name_base + '2c')(x)
shortcut = Convolution2D(nb_filter3,
1,
1,
subsample=strides,
name=conv_name_base + '1',
bias=False)(input_tensor)
shortcut = BatchNormalization(epsilon=eps,
axis=bn_axis,
name=bn_name_base + '1')(shortcut)
shortcut = Scale(axis=bn_axis, name=scale_name_base + '1')(shortcut)
x = merge([x, shortcut], mode='sum', name='res' + str(stage) + block)
x = Activation('relu', name='res' + str(stage) + block + '_relu')(x)
return x
def IC(input, p):
eps = 1.1e-5
x = BatchNormalization(epsilon=eps, axis=bn_axis)(input)
x = Scale(axis=bn_axis)(x)
x = Dropout(p)(x)
return x