ECA-Net通道注意力机制复现

ECA-ResNet50花朵识别

.gitignore

@@ -36,3 +36,4 @@
 /经典网络/ShuffleNet/checkpoint/
 /经典网络/ShuffleNet/checkpoint_v2/
 /经典网络/ResNext/checkpoint/
+/经典网络/ECANet/checkpoint/

循环神经网络/Attention/图像中的注意力机制/ECANet.py

 import math
 
 import tensorflow as tf
-from keras import backend as K
 from keras.layers import (Activation, Add, Concatenate, Conv1D, Conv2D, Dense,
                           GlobalAveragePooling2D, GlobalMaxPooling2D, Lambda, BatchNormalization,
                           Reshape, multiply,Input)
@@ -10,16 +9,18 @@ from keras.models import Model
 ECA模块的思想是非常简单的，它去除了原来SE模块中的全连接层，
 直接在全局平均池化之后的特征上通过一个1D卷积进行学习。
 '''
-def eca_block(input_feature, b=1, gamma=2, name=""):
-    channel = input_feature._keras_shape[-1]
+def ECA_Block(input_feature, b=1, gamma=2, name=""):
+    channel = input_feature.shape[-1]
+    # 根据公式计算自适应卷积核大小
     kernel_size = int(abs((math.log(channel, 2) + b) / gamma))
+    # 如果是kernel_size是偶数，就使用，否则变成偶数
     kernel_size = kernel_size if kernel_size % 2 else kernel_size + 1
     # [c]
-    avg_pool = GlobalAveragePooling2D()(input_feature)
+    x = GlobalAveragePooling2D()(input_feature)
     # [c,1]
-    x = Reshape((-1, 1))(avg_pool)
+    x = Reshape((-1, 1))(x)
     #
-    x = Conv1D(1, kernel_size=kernel_size, padding="same", name="eca_layer_" + str(name), use_bias=False, )(x)
+    x = Conv1D(1, kernel_size=kernel_size, padding="same",use_bias=False,name="eca_layer_" + str(name))(x)
     x = Activation('sigmoid')(x)
     # [c,1]=>[1,1,c]
     x = Reshape((1, 1, -1))(x)
@@ -29,6 +30,6 @@ def eca_block(input_feature, b=1, gamma=2, name=""):
 
 if __name__ == '__main__':
     inputs=Input([26,26,512])
-    x=eca_block(inputs)
+    x=ECA_Block(inputs)
     model=Model(inputs,x)
     model.summary()
\ No newline at end of file

经典网络/ECANet/ECANet-block.py

+import tensorflow as tf
+import math
+from tensorflow.keras.layers import GlobalAveragePooling2D,Conv1D,Reshape,Activation,multiply
+from tensorflow.keras.layers import Input
+from tensorflow.keras.models import Model
+from plot_model import plot_model
+
+
+'''
+ECA模块的思想是非常简单的，它去除了原来SE模块中的全连接层，
+直接在全局平均池化之后的特征上通过一个1D卷积进行学习。
+'''
+def eca_block(input_feature, b=1, gamma=2, name=""):
+    channel = input_feature.shape[-1]
+    # 根据公式计算自适应卷积核大小
+    kernel_size = int(abs((math.log(channel, 2) + b) / gamma))
+    # 如果是kernel_size是偶数，就使用，否则变成偶数
+    kernel_size = kernel_size if kernel_size % 2 else kernel_size + 1
+    # [c]
+    x = GlobalAveragePooling2D()(input_feature)
+    # [c,1]
+    x = Reshape((-1, 1))(x)
+    #
+    x = Conv1D(1, kernel_size=kernel_size, padding="same",use_bias=False,name="eca_layer_" + str(name))(x)
+    x = Activation('sigmoid')(x)
+    # [c,1]=>[1,1,c]
+    x = Reshape((1, 1, -1))(x)
+
+    output = multiply([input_feature, x])
+    return output
+
+if __name__ == '__main__':
+    inputs=Input([26,25,512])
+    x=eca_block(inputs)
+    model=Model(inputs,x)
+    model.summary()
+    plot_model(model,to_file='img/ECANet-block.png')
\ No newline at end of file