File size: 9,762 Bytes
edcf5ee |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 |
"""
attention modules in ['SimAM', 'CBAM', 'SE', 'GAM'] were applied in the ablation study
ver： Dec 24th 15：00
ref:
https://github.com/xmu-xiaoma666/External-Attention-pytorch
"""
import torch
import torch.nn as nn
import math
import torch.nn.functional as F
from torch.nn import init
# help func
class BasicConv(nn.Module):
def __init__(self, in_planes, out_planes, kernel_size, stride=1, padding=0, dilation=1, groups=1, relu=True,
bn=True, bias=False):
super(BasicConv, self).__init__()
self.out_channels = out_planes
self.conv = nn.Conv2d(in_planes, out_planes, kernel_size=kernel_size, stride=stride, padding=padding,
dilation=dilation, groups=groups, bias=bias)
self.bn = nn.BatchNorm2d(out_planes, eps=1e-5, momentum=0.01, affine=True) if bn else None
self.relu = nn.ReLU() if relu else None
def forward(self, x):
x = self.conv(x)
if self.bn is not None:
x = self.bn(x)
if self.relu is not None:
x = self.relu(x)
return x
class Flatten(nn.Module):
def forward(self, x):
return x.view(x.size(0), -1)
class ChannelGate(nn.Module):
def __init__(self, gate_channels, reduction_ratio=16, pool_types=['avg', 'max']):
super(ChannelGate, self).__init__()
self.gate_channels = gate_channels
self.mlp = nn.Sequential(
Flatten(),
nn.Linear(gate_channels, int(gate_channels // reduction_ratio)),
nn.ReLU(),
nn.Linear(int(gate_channels // reduction_ratio), gate_channels)
)
self.pool_types = pool_types
def forward(self, x):
channel_att_sum = None
for pool_type in self.pool_types:
if pool_type == 'avg':
avg_pool = F.avg_pool2d(x, (x.size(2), x.size(3)), stride=(x.size(2), x.size(3)))
channel_att_raw = self.mlp(avg_pool)
elif pool_type == 'max':
max_pool = F.max_pool2d(x, (x.size(2), x.size(3)), stride=(x.size(2), x.size(3)))
channel_att_raw = self.mlp(max_pool)
elif pool_type == 'lp':
lp_pool = F.lp_pool2d(x, 2, (x.size(2), x.size(3)), stride=(x.size(2), x.size(3)))
channel_att_raw = self.mlp(lp_pool)
elif pool_type == 'lse':
# LSE pool only
lse_pool = logsumexp_2d(x)
channel_att_raw = self.mlp(lse_pool)
if channel_att_sum is None:
channel_att_sum = channel_att_raw
else:
channel_att_sum = channel_att_sum + channel_att_raw
scale = F.sigmoid(channel_att_sum).unsqueeze(2).unsqueeze(3).expand_as(x)
return x * scale
def logsumexp_2d(tensor):
tensor_flatten = tensor.view(tensor.size(0), tensor.size(1), -1)
s, _ = torch.max(tensor_flatten, dim=2, keepdim=True)
outputs = s + (tensor_flatten - s).exp().sum(dim=2, keepdim=True).log()
return outputs
class ChannelPool(nn.Module):
def forward(self, x):
return torch.cat((torch.max(x, 1)[0].unsqueeze(1), torch.mean(x, 1).unsqueeze(1)), dim=1)
class SpatialGate(nn.Module):
def __init__(self):
super(SpatialGate, self).__init__()
kernel_size = 7
self.compress = ChannelPool()
self.spatial = BasicConv(2, 1, kernel_size, stride=1, padding=int((kernel_size - 1) // 2), relu=False)
def forward(self, x):
x_compress = self.compress(x)
x_out = self.spatial(x_compress)
scale = F.sigmoid(x_out) # broadcasting
return x * scale
# attention modules：
class cbam_module(nn.Module):
"""
module：CBAM
input、output= b, c, h, w
paper：
https://arxiv.org/abs/1807.06521
code：
https://github.com/ZjjConan/SimAM/blob/master/networks/attentions
"""
def __init__(self, gate_channels, reduction=16, pool_types=['avg', 'max'], no_spatial=False):
super(cbam_module, self).__init__()
self.ChannelGate = ChannelGate(gate_channels, reduction, pool_types)
self.no_spatial = no_spatial
if not no_spatial:
self.SpatialGate = SpatialGate()
@staticmethod
def get_module_name():
return "cbam"
def forward(self, x):
x_out = self.ChannelGate(x)
if not self.no_spatial:
x_out = self.SpatialGate(x_out)
return x_out
class se_module(nn.Module):
"""
module: SE
input、output= b, c, h, w
from paper Squeeze-and-Excitation Networks
SE-Net https://arxiv.org/abs/1709.01507
code：
https://github.com/ZjjConan/SimAM/blob/master/networks/attentions
"""
def __init__(self, channel, reduction=16):
super(se_module, self).__init__()
self.avg_pool = nn.AdaptiveAvgPool2d(1)
self.fc = nn.Sequential(
nn.Linear(channel, int(channel // reduction), bias=False),
nn.ReLU(inplace=True),
nn.Linear(int(channel // reduction), channel, bias=False),
nn.Sigmoid()
)
@staticmethod
def get_module_name():
return "se"
def forward(self, x):
b, c, _, _ = x.size()
y = self.avg_pool(x).view(b, c)
y = self.fc(y).view(b, c, 1, 1)
return x * y
class simam_module(torch.nn.Module):
"""
module:SimAM
input、output= b, c, h, w
paper:(ICML)
SimAM: A Simple, Parameter-Free Attention Module for Convolutional Neural Networks
code:
https://github.com/ZjjConan/SimAM/blob/master/networks/attentions/simam_module.py
"""
def __init__(self, channels=None, e_lambda=1e-4):
super(simam_module, self).__init__()
self.activaton = nn.Sigmoid()
self.e_lambda = e_lambda
def __repr__(self):
s = self.__class__.__name__ + '('
s += ('lambda=%f)' % self.e_lambda)
return s
@staticmethod
def get_module_name():
return "simam"
def forward(self, x):
b, c, h, w = x.size()
n = w * h - 1
x_minus_mu_square = (x - x.mean(dim=[2, 3], keepdim=True)).pow(2)
y = x_minus_mu_square / (4 * (x_minus_mu_square.sum(dim=[2, 3], keepdim=True) / n + self.e_lambda)) + 0.5
return x * self.activaton(y)
class ResidualAttention(nn.Module):
"""
module: ResidualAttention
input、output= b, c, h, w
Paper：ICCV 2021 Residual Attention: A Simple but Effective Method for Multi-Label Recognition
code：https://github.com/xmu-xiaoma666/External-Attention-pytorch/blob/master/attention/ResidualAttention.py
"""
def __init__(self, channel=512, num_class=1000, la=0.2):
super().__init__()
self.la = la
self.fc = nn.Conv2d(in_channels=channel, out_channels=num_class, kernel_size=1, stride=1, bias=False)
def forward(self, x):
b, c, h, w = x.shape
y_raw = self.fc(x).flatten(2) # b,num_class,hxw
y_avg = torch.mean(y_raw, dim=2) # b,num_class
y_max = torch.max(y_raw, dim=2)[0] # b,num_class
score = y_avg + self.la * y_max
return score
class eca_module(nn.Module):
"""Constructs a ECA module.
Args:
channel: Number of channels of the input feature map
k_size: Adaptive selection of kernel size
"""
def __init__(self, channel, k_size=3):
super(eca_module, self).__init__()
self.avg_pool = nn.AdaptiveAvgPool2d(1)
self.conv = nn.Conv1d(1, 1, kernel_size=k_size, padding=(k_size - 1) // 2, bias=False)
self.sigmoid = nn.Sigmoid()
def forward(self, x):
# x: input features with shape [b, c, h, w]
b, c, h, w = x.size()
# feature descriptor on the global spatial information
y = self.avg_pool(x)
# Two different branches of ECA module
y = self.conv(y.squeeze(-1).transpose(-1, -2)).transpose(-1, -2).unsqueeze(-1)
# Multi-scale information fusion
y = self.sigmoid(y)
return x * y.expand_as(x)
class GAM_Attention(nn.Module):
"""
module：GAM
input= b, in_channels, h, w
output= b, out_channels, h, w
paper：
Global Attention Mechanism: Retain Information to Enhance Channel-Spatial Interactions
https://arxiv.org/abs/2112.05561
code：
https://mp.weixin.qq.com/s/VL6rXjyUDmHToYTqM32hUg
"""
def __init__(self, in_channels, out_channels, rate=4):
super(GAM_Attention, self).__init__()
self.channel_attention = nn.Sequential(
nn.Linear(in_channels, int(in_channels / rate)),
nn.ReLU(inplace=True),
nn.Linear(int(in_channels / rate), in_channels)
)
self.spatial_attention = nn.Sequential(
nn.Conv2d(in_channels, int(in_channels / rate), kernel_size=7, padding=3),
nn.BatchNorm2d(int(in_channels / rate)),
nn.ReLU(inplace=True),
nn.Conv2d(int(in_channels / rate), out_channels, kernel_size=7, padding=3),
nn.BatchNorm2d(out_channels)
)
def forward(self, x):
b, c, h, w = x.shape
x_permute = x.permute(0, 2, 3, 1).view(b, -1, c)
x_att_permute = self.channel_attention(x_permute).view(b, h, w, c)
x_channel_att = x_att_permute.permute(0, 3, 1, 2)
x = x * x_channel_att
x_spatial_att = self.spatial_attention(x).sigmoid()
out = x * x_spatial_att
return out
|