| import torch |
| import torch.nn as nn |
| import torch.nn.functional as F |
|
|
| class ASPPModule(nn.Module): |
| def __init__(self, inplanes, planes, kernel_size, padding, dilation, norm_fn=None): |
| super().__init__() |
| self.atrous_conv = nn.Conv2d(inplanes, planes, kernel_size=kernel_size, stride=1, padding=padding, dilation=dilation, bias=False) |
| self.bn = norm_fn(planes) |
| self.relu = nn.ReLU(inplace=True) |
|
|
| self.initialize([self.atrous_conv, self.bn]) |
|
|
| def forward(self, x): |
| x = self.atrous_conv(x) |
| x = self.bn(x) |
| return self.relu(x) |
|
|
| def initialize(self, modules): |
| for m in modules: |
| if isinstance(m, nn.Conv2d): |
| torch.nn.init.kaiming_normal_(m.weight) |
| elif isinstance(m, nn.BatchNorm2d): |
| m.weight.data.fill_(1) |
| m.bias.data.zero_() |
|
|
| class ASPP(nn.Module): |
| def __init__(self, output_stride, norm_fn, inplanes=2048): |
| super().__init__() |
|
|
| inplanes = inplanes |
|
|
| if output_stride == 16: |
| dilations = [1, 6, 12, 18] |
| elif output_stride == 8: |
| dilations = [1, 12, 24, 36] |
| |
| self.aspp1 = ASPPModule(inplanes, 256, 1, padding=0, dilation=dilations[0], norm_fn=norm_fn) |
| self.aspp2 = ASPPModule(inplanes, 256, 3, padding=dilations[1], dilation=dilations[1], norm_fn=norm_fn) |
| self.aspp3 = ASPPModule(inplanes, 256, 3, padding=dilations[2], dilation=dilations[2], norm_fn=norm_fn) |
| self.aspp4 = ASPPModule(inplanes, 256, 3, padding=dilations[3], dilation=dilations[3], norm_fn=norm_fn) |
|
|
| self.global_avg_pool = nn.Sequential( |
| nn.AdaptiveAvgPool2d((1, 1)), |
| nn.Conv2d(inplanes, 256, 1, stride=1, bias=False), |
| norm_fn(256), |
| nn.ReLU(inplace=True), |
| ) |
|
|
| self.conv1 = nn.Conv2d(1280, 256, 1, bias=False) |
| self.bn1 = norm_fn(256) |
| self.relu = nn.ReLU(inplace=True) |
| self.dropout = nn.Dropout(0.5) |
| |
| self.initialize([self.conv1, self.bn1] + list(self.global_avg_pool.modules())) |
| |
| def forward(self, x): |
| x1 = self.aspp1(x) |
| x2 = self.aspp2(x) |
| x3 = self.aspp3(x) |
| x4 = self.aspp4(x) |
|
|
| x5 = self.global_avg_pool(x) |
| x5 = F.interpolate(x5, size=x4.size()[2:], mode='bilinear', align_corners=True) |
|
|
| x = torch.cat((x1, x2, x3, x4, x5), dim=1) |
|
|
| x = self.conv1(x) |
| x = self.bn1(x) |
| x = self.relu(x) |
| x = self.dropout(x) |
|
|
| return x |
|
|
| def initialize(self, modules): |
| for m in modules: |
| if isinstance(m, nn.Conv2d): |
| torch.nn.init.kaiming_normal_(m.weight) |
| elif isinstance(m, nn.BatchNorm2d): |
| m.weight.data.fill_(1) |
| m.bias.data.zero_() |
|
|
| class Decoder(nn.Module): |
| def __init__(self, num_classes, low_level_inplanes, norm_fn, kernel_size=3, padding=1): |
| super().__init__() |
| |
| self.kernel_size = kernel_size |
| self.padding = padding |
|
|
| self.conv1 = nn.Conv2d(low_level_inplanes, 48, 1, bias=False) |
| self.bn1 = norm_fn(48) |
| self.relu = nn.ReLU(inplace=True) |
| |
| self.classifier = nn.Sequential( |
| nn.Conv2d(304, 256, kernel_size=self.kernel_size, stride=1, padding=self.padding, bias=False), |
| norm_fn(256), |
| nn.ReLU(inplace=True), |
| nn.Dropout(0.5), |
|
|
| nn.Conv2d(256, 256, kernel_size=self.kernel_size, stride=1, padding=self.padding, bias=False), |
| norm_fn(256), |
| nn.ReLU(inplace=True), |
| nn.Dropout(0.1), |
| nn.Conv2d(256, num_classes, kernel_size=1, stride=1) |
| ) |
|
|
| self.initialize([self.conv1, self.bn1] + list(self.classifier.modules())) |
|
|
| def forward(self, x, x_low_level): |
| x_low_level = self.conv1(x_low_level) |
| x_low_level = self.bn1(x_low_level) |
| x_low_level = self.relu(x_low_level) |
|
|
| x = F.interpolate(x, size=x_low_level.size()[2:], mode='bilinear', align_corners=True) |
| x = torch.cat((x, x_low_level), dim=1) |
| x = self.classifier(x) |
|
|
| return x |
|
|
| def initialize(self, modules): |
| for m in modules: |
| if isinstance(m, nn.Conv2d): |
| torch.nn.init.kaiming_normal_(m.weight) |
| elif isinstance(m, nn.BatchNorm2d): |
| m.weight.data.fill_(1) |
| m.bias.data.zero_() |
|
|
| class Decoder_Attention(nn.Module): |
| def __init__(self, num_classes, low_level_inplanes, norm_fn, kernel_size=3, padding=1, attention_mode='CBAM'): |
| super().__init__() |
| |
| self.kernel_size = kernel_size |
| self.padding = padding |
|
|
| self.conv1 = nn.Conv2d(low_level_inplanes, 48, 1, bias=False) |
| self.bn1 = norm_fn(48) |
| self.relu = nn.ReLU(inplace=True) |
| |
| self.attention_mode = attention_mode |
|
|
| if attention_mode == 'CBAM': |
| self.attention = CBAM(304) |
| self.attention1 = CBAM(256) |
| elif attention_mode == 'SA': |
| self.attention = SpatialAttention() |
| self.attention1 = SpatialAttention() |
| else: |
| self.attention = ChannelAttention(304) |
| self.attention1 = ChannelAttention(256) |
|
|
| self.conv2 = nn.Conv2d(304, 256, kernel_size=self.kernel_size, stride=1, padding=self.padding, bias=False) |
| self.bn2 = norm_fn(256) |
| self.dropout2 = nn.Dropout(0.5) |
|
|
| self.conv3 = nn.Conv2d(256, 256, kernel_size=self.kernel_size, stride=1, padding=self.padding, bias=False) |
| self.bn3 = norm_fn(256) |
| self.dropout3 = nn.Dropout(0.1) |
| self.conv4 = nn.Conv2d(256, num_classes, kernel_size=1, stride=1) |
|
|
| self.initialize([self.conv1, self.bn1, self.conv2, self.bn2, self.conv3, self.bn3, self.conv4, self.attention, self.attention1]) |
|
|
| def forward(self, x, x_low_level): |
| x_low_level = self.conv1(x_low_level) |
| x_low_level = self.bn1(x_low_level) |
| x_low_level = self.relu(x_low_level) |
|
|
| x = F.interpolate(x, size=x_low_level.size()[2:], mode='bilinear', align_corners=True) |
| x = torch.cat((x, x_low_level), dim=1) |
| x = self.attention(x) |
| x = self.conv2(x) |
| x = self.bn2(x) |
| x = self.attention1(x) |
| x = self.relu(x) |
| x = self.dropout2(x) |
| x = self.conv3(x) |
| x = self.relu(x) |
| x = self.dropout3(x) |
| x = self.conv4(x) |
| return x |
|
|
| def initialize(self, modules): |
| for m in modules: |
| if isinstance(m, nn.Conv2d): |
| torch.nn.init.kaiming_normal_(m.weight) |
| elif isinstance(m, nn.BatchNorm2d): |
| m.weight.data.fill_(1) |
| m.bias.data.zero_() |
|
|
| class ChannelAttention(nn.Module): |
| def __init__(self, in_planes, ratio=16): |
| super(ChannelAttention, self).__init__() |
| self.avg_pool = nn.AdaptiveAvgPool2d(1) |
| self.max_pool = nn.AdaptiveMaxPool2d(1) |
|
|
| self.fc1 = nn.Conv2d(in_planes, in_planes // ratio, 1, bias=False) |
| self.relu1 = nn.ReLU() |
| self.fc2 = nn.Conv2d(in_planes // ratio, in_planes, 1, bias=False) |
| self.sigmoid = nn.Sigmoid() |
|
|
| def forward(self, x): |
| avg_out = self.fc2(self.relu1(self.fc1(self.avg_pool(x)))) |
| max_out = self.fc2(self.relu1(self.fc1(self.max_pool(x)))) |
| out = avg_out + max_out |
| return x*self.sigmoid(out) |
|
|
| class SpatialAttention(nn.Module): |
| def __init__(self, kernel_size=7): |
| super(SpatialAttention, self).__init__() |
|
|
| assert kernel_size in (3, 7), 'kernel size must be 3 or 7' |
| padding = 3 if kernel_size == 7 else 1 |
|
|
| self.conv1 = nn.Conv2d(2, 1, kernel_size, padding=padding, bias=False) |
| self.sigmoid = nn.Sigmoid() |
|
|
| def forward(self, x): |
| avg_out = torch.mean(x, dim=1, keepdim=True) |
| max_out, _ = torch.max(x, dim=1, keepdim=True) |
| y = torch.cat([avg_out, max_out], dim=1) |
| y = self.conv1(y) |
| return x*self.sigmoid(y) |
|
|
| class CBAM(nn.Module): |
| def __init__(self, in_planes, ratio=16, kernel_size=7): |
| super(CBAM, self).__init__() |
| self.ca = ChannelAttention(in_planes, ratio) |
| self.sa = SpatialAttention(kernel_size) |
| |
| def forward(self, x): |
| out = self.ca(x) |
| result = self.sa(out) |
| return result |
|
|
|
|
|
|
