肺结节识别系列之问题解决一
在进行项目复现时，遇到了up6 = merge([UpSampling2D(size=(2, 2))(conv5), conv4], mode=‘concat’, concat_axis=1) TypeError: ‘module’ object is not callable 错误
百度了许多方案都没有解决，通过看官方文档解决了问题，以下解决方法及codes
编译环境：Anaconda3、Python3.7
- 源代码
def unet_model():
inputs = Input((1,512, 512))
conv1 = Conv2D(width, (3, 3), activation='relu', padding='same')(inputs)
conv1 = BatchNormalization(axis = 1)(conv1)
conv1 = Conv2D(width, (3, 3), activation='relu', padding='same')(conv1)
pool1 = MaxPooling2D(pool_size=(2, 2))(conv1)
conv2 = Conv2D(width*2, (3, 3), activation='relu', padding='same')(pool1)
conv2 = BatchNormalization(axis = 1)(conv2)
conv2 = Conv2D(width*2, (3, 3), activation='relu', padding='same')(conv2)
pool2 = MaxPooling2D(pool_size=(2, 2))(conv2)
conv3 = Conv2D(width*4, (3, 3), activation='relu', padding='same')(pool2)
conv3 = BatchNormalization(axis = 1)(conv3)
conv3 = Conv2D(width*4, (3, 3), activation='relu', padding='same')(conv3)
pool3 = MaxPooling2D(pool_size=(2, 2))(conv3)
conv4 = Conv2D(width*8, (3, 3), activation='relu', padding='same')(pool3)
conv4 = BatchNormalization(axis = 1)(conv4)
conv4 = Conv2D(width*8, (3, 3), activation='relu', padding='same')(conv4)
pool4 = MaxPooling2D(pool_size=(2, 2))(conv4)
conv5 = Conv2D(width*16, (3, 3), activation='relu', padding='same')(pool4)
conv5 = BatchNormalization(axis = 1)(conv5)
conv5 = Conv2D(width*16, (3, 3), activation='relu', padding='same')(conv5)
up6 = merge([UpSampling2D(size=(2, 2))(conv5), conv4], mode='concat', concat_axis=1)
conv6 = SpatialDropout2D(0.35)(up6)
conv6 = Conv2D(width*8, (3, 3), activation='relu', padding='same')(conv6)
conv6 = Conv2D(width*8, (3, 3), activation='relu', padding='same')(conv6)
up7 = merge([UpSampling2D(size=(2, 2))(conv6), conv3], mode='concat', concat_axis=1)
conv7 = SpatialDropout2D(0.35)(up7)
conv7 = Conv2D(width*4, (3, 3), activation='relu', padding='same')(conv7)
conv7 = Conv2D(width*4, (3, 3), activation='relu', padding='same')(conv7)
up8 = merge([UpSampling2D(size=(2, 2))(conv7), conv2], mode='concat', concat_axis=1)
conv8 = SpatialDropout2D(0.35)(up8)
conv8 = Conv2D(width*2, (3, 3), activation='relu', padding='same')(conv8)
conv8 = Conv2D(width*2, (3, 3), activation='relu', padding='same')(conv8)
up9 = merge([UpSampling2D(size=(2, 2))(conv8), conv1], mode='concat', concat_axis=1)
conv9 = SpatialDropout2D(0.35)(up9)
conv9 = Conv2D(width, (3, 3), activation='relu', padding='same')(conv9)
conv9 = Conv2D(width, (3, 3), activation='relu', padding='same')(conv9)
conv10 = Conv2D(1, (1, 1), activation='sigmoid')(conv9)
model = Model(input=inputs, output=conv10)
model.compile(optimizer=Adam(lr=1e-5), loss=dice_coef_loss, metrics=[dice_coef])
return model
查阅参考了Keras的官方文档融合层(Merge)和卷积层(Convolutional)
卷积层的UpSampling2D
keras.layers.UpSampling2D(size=(2, 2), data_format=None, interpolation=‘nearest’)
融合层的concatenate
keras.layers.concatenate(inputs, axis=-1)
将codes修改为
def unet_model():
inputs = Input((1,512, 512))
conv1 = Conv2D(width, 3, activation='relu', padding='same')(inputs)
conv1 = BatchNormalization(axis = 1)(conv1)
conv1 = Conv2D(width, 3, activation='relu', padding='same')(conv1)
pool1 = MaxPooling2D(pool_size=(2, 2))(conv1)
conv2 = Conv2D(width*2, 3, activation='relu', padding='same')(pool1)
conv2 = BatchNormalization(axis = 1)(conv2)
conv2 = Conv2D(width*2, 3, activation='relu', padding='same')(conv2)
pool2 = MaxPooling2D(pool_size=(2, 2))(conv2)
conv3 = Conv2D(width*4, 3, activation='relu', padding='same')(pool2)
conv3 = BatchNormalization(axis = 1)(conv3)
conv3 = Conv2D(width*4, 3, activation='relu', padding='same')(conv3)
pool3 = MaxPooling2D(pool_size=(2, 2))(conv3)
conv4 = Conv2D(width*8, 3, activation='relu', padding='same')(pool3)
conv4 = BatchNormalization(axis = 1)(conv4)
conv4 = Conv2D(width*8, 3, activation='relu', padding='same')(conv4)
pool4 = MaxPooling2D(pool_size=(2, 2))(conv4)
conv5 = Conv2D(width*16, 3, activation='relu', padding='same')(pool4)
conv5 = BatchNormalization(axis = 1)(conv5)
conv5 = Conv2D(width*16, 3, activation='relu', padding='same')(conv5)
up6 = UpSampling2D(size=(2, 2))(conv5)
up6 = concatenate([up6, conv4], axis=1)
conv6 = SpatialDropout2D(0.35)(up6)
conv6 = Conv2D(width*8, 3, activation='relu', padding='same')(conv6)
conv6 = Conv2D(width*8, 3, activation='relu', padding='same')(conv6)
up7 = UpSampling2D(size=(2, 2))(conv6)
up7 = concatenate([up7, conv3], axis=1)
conv7 = SpatialDropout2D(0.35)(up7)
conv7 = Conv2D(width*4, 3, activation='relu', padding='same')(conv7)
conv7 = Conv2D(width*4, 3, activation='relu', padding='same')(conv7)
up8 = UpSampling2D(size=(2, 2))(conv7)
up8 = concatenate([up8, conv2], axis=1)
conv8 = SpatialDropout2D(0.35)(up8)
conv8 = Conv2D(width*2, 3, activation='relu', padding='same')(conv8)
conv8 = Conv2D(width*2, 3, activation='relu', padding='same')(conv8)
up9 = UpSampling2D(size=(2, 2))(conv8)
up9 = concatenate([up9, conv1], axis=1)
conv9 = SpatialDropout2D(0.35)(up9)
conv9 = Conv2D(width, 3, activation='relu', padding='same')(conv9)
conv9 = Conv2D(width, 3, activation='relu', padding='same')(conv9)
conv10 = Conv2D(1, 1, activation='sigmoid')(conv9)
model = Model(input=inputs, output=conv10)
model.compile(optimizer=Adam(lr=1e-5), loss=dice_coef_loss, metrics=[dice_coef])
return model
成功解决问题
注：在调用concatenate时，记得添加from keras.layers import concatenate
26/4/2019 21:38
又发现了更简洁的代码up6 = concatenate([UpSampling2D(size=(2, 2))(conv5), conv4], axis=1)
按上面修改即可成功。
