【论文解读】用于卷积神经网络的注意力机制(Attention)----CBAM: Convolutional Block Attention Module

论文：CBAM: Convolutional Block Attention Module
收录于：ECCV 2018

摘要

论文提出了Convolutional Block Attention Module(CBAM)，这是一种为卷积神将网络设计的，简单有效的注意力模块(Attention Module)。对于卷积神经网络生成的feature map，CBAM从通道和空间两个维度计算feature map的attention map，然后将attention map与输入的feature map相乘来进行特征的自适应学习。CBAM是一个轻量的通用模块，可以将其融入到各种卷积神经网络中进行端到端的训练。

主要思想

对于一个中间层的feature map：$F\in {\mathbb{R}}^{C\ast H\ast W}$，CBAM将会顺序推理出1维的channel attention map $M_c\in {\mathbb{R}}^{C\ast 1\ast 1}$以及2维的spatial attention map $M_s\in {\mathbb{R}}^{1\ast H\ast W}$，整个过程如下所示：
$$F^{{}^{\prime }}=M_c(F)\otimes F$$
$$F^{{}^{\prime \prime }}=M_s(F^{{}^{\prime }})\otimes F^{{}^{\prime }}$$
其中$\otimes$为element-wise multiplication，首先将channel attention map与输入的feature map相乘得到$F^{{}^{\prime }}$，之后计算$F^{{}^{\prime }}$的spatial attention map，并将两者相乘得到最终的输出$F^{{}^{\prime \prime }}$。下图为CBAM的示意图：

Channel attention module

feature map 的每个channel都被视为一个feature detector，channel attention主要关注于输入图片中什么(what)是有意义的。为了高效地计算channel attention，论文使用最大池化和平均池化对feature map在空间维度上进行压缩，得到两个不同的空间背景描述：$F_{max}^c$和$F_{avg}^c$。使用由MLP组成的共享网络对这两个不同的空间背景描述进行计算得到channel attention map：$M_c\in {\mathbb{R}}^{C\ast 1\ast 1}$。计算过程如下：
$$M_c(F)=\sigma (MLP(AvgPool(F)))+\sigma (MLP(MaxPool(F)))$$
$$M_c(F)=\sigma (W_1(W_0(F_{avg}^c)))+\sigma (W_1(W_0(F_{max}^c)))$$

其中$W_0\in {\mathbb{R}}^{C/r\ast C}$，$W_1\in {\mathbb{R}}^{C\ast C/r}$，$W_0$后使用了Relu作为激活函数。

Spatial attention module.

与channel attention不同，spatial attention主要关注于位置信息(where)。为了计算spatial attention，论文首先在channel的维度上使用最大池化和平均池化得到两个不同的特征描述$F_{max}^s\in {\mathbb{R}}_{1\ast H\ast W}$和$F_{avg}^s\in {\mathbb{R}}_{1\ast H\ast W}$，然后使用concatenation将两个特征描述合并，并使用卷积操作生成spatial attention map $M_s(F)\in {\mathbb{R}}_{H\ast W}$。计算过程如下：
$$M_s(F)=\sigma (f^{7\ast 7}([AvgPool(F);MaxPool(F)]))$$
$$M_s(F)=\sigma (f^{7\ast 7}([F_{avg}^s;F_{max}^s]))$$
其中，$f^{7\ast 7}$表示7*7的卷积层

下图为channel attention和spatial attention的示意图：

代码

环境：tensorflow 1.9

"""
@Time   : 2018/10/19
@Author : Li YongHong
@Email  : [email protected]
@File   : test.py
"""
import tensorflow as tf
import numpy as np

slim = tf.contrib.slim

def combined_static_and_dynamic_shape(tensor):
  """Returns a list containing static and dynamic values for the dimensions.

  Returns a list of static and dynamic values for shape dimensions. This is
  useful to preserve static shapes when available in reshape operation.

  Args:
    tensor: A tensor of any type.

  Returns:
    A list of size tensor.shape.ndims containing integers or a scalar tensor.
  """
  static_tensor_shape = tensor.shape.as_list()
  dynamic_tensor_shape = tf.shape(tensor)
  combined_shape = []
  for index, dim in enumerate(static_tensor_shape):
    if dim is not None:
      combined_shape.append(dim)
    else:
      combined_shape.append(dynamic_tensor_shape[index])
  return combined_shape

def convolutional_block_attention_module(feature_map, index, inner_units_ratio=0.5):
    """
    CBAM: convolution block attention module, which is described in "CBAM: Convolutional Block Attention Module"
    Architecture : "https://arxiv.org/pdf/1807.06521.pdf"
    If you want to use this module, just plug this module into your network
    :param feature_map : input feature map
    :param index : the index of convolution block attention module
    :param inner_units_ratio: output units number of fully connected layer: inner_units_ratio*feature_map_channel
    :return:feature map with channel and spatial attention
    """
    with tf.variable_scope("cbam_%s" % (index)):
        feature_map_shape = combined_static_and_dynamic_shape(feature_map)
        # channel attention
        channel_avg_weights = tf.nn.avg_pool(
            value=feature_map,
            ksize=[1, feature_map_shape[1], feature_map_shape[2], 1],
            strides=[1, 1, 1, 1],
            padding='VALID'
        )
        channel_max_weights = tf.nn.max_pool(
            value=feature_map,
            ksize=[1, feature_map_shape[1], feature_map_shape[2], 1],
            strides=[1, 1, 1, 1],
            padding='VALID'
        )
        channel_avg_reshape = tf.reshape(channel_avg_weights,
                                         [feature_map_shape[0], 1, feature_map_shape[3]])
        channel_max_reshape = tf.reshape(channel_max_weights,
                                         [feature_map_shape[0], 1, feature_map_shape[3]])
        channel_w_reshape = tf.concat([channel_avg_reshape, channel_max_reshape], axis=1)

        fc_1 = tf.layers.dense(
            inputs=channel_w_reshape,
            units=feature_map_shape[3] * inner_units_ratio,
            name="fc_1",
            activation=tf.nn.relu
        )
        fc_2 = tf.layers.dense(
            inputs=fc_1,
            units=feature_map_shape[3],
            name="fc_2",
            activation=tf.nn.sigmoid
        )
        channel_attention = tf.reduce_sum(fc_2, axis=1, name="channel_attention_sum")
        channel_attention = tf.reshape(channel_attention, shape=[feature_map_shape[0], 1, 1, feature_map_shape[3]])
        feature_map_with_channel_attention = tf.multiply(feature_map, channel_attention)
        # spatial attention
        channel_wise_avg_pooling = tf.reduce_mean(feature_map_with_channel_attention, axis=3)
        channel_wise_max_pooling = tf.reduce_max(feature_map_with_channel_attention, axis=3)

        channel_wise_avg_pooling = tf.reshape(channel_wise_avg_pooling,
                                              shape=[feature_map_shape[0], feature_map_shape[1], feature_map_shape[2],
                                                     1])
        channel_wise_max_pooling = tf.reshape(channel_wise_max_pooling,
                                              shape=[feature_map_shape[0], feature_map_shape[1], feature_map_shape[2],
                                                     1])

        channel_wise_pooling = tf.concat([channel_wise_avg_pooling, channel_wise_max_pooling], axis=3)
        spatial_attention = slim.conv2d(
            channel_wise_pooling,
            1,
            [3, 3],
            padding='SAME',
            activation_fn=tf.nn.sigmoid,
            scope="spatial_attention_conv"
        )
        feature_map_with_attention = tf.multiply(feature_map_with_channel_attention, spatial_attention)
        return feature_map_with_attention

#example
feature_map = tf.constant(np.random.rand(2,8,8,32), dtype=tf.float16)
feature_map_with_attention = convolutional_block_attention_module(feature_map)

with tf.Session() as sess:
    init = tf.global_variables_initializer()
    sess.run(init)
    result = sess.run(feature_map_with_attention)
    print(result.shape)