每日Attention学习18——Grouped Attention Gate

模块出处

[ICLR 25 Submission] [link] UltraLightUNet: Rethinking U-shaped Network with Multi-kernel Lightweight Convolutions for Medical Image Segmentation

---

模块名称

Grouped Attention Gate (GAG)

---

模块作用

轻量特征融合

---

模块结构

---

模块特点

• 特征融合前使用Group Conv进行处理，比标准卷积更加轻量
• 将融合得到的粗特征视为Spatial Attention Map, 并与Encoder特征相乘，从而实现名字中"Gate"的效果
• 相较于特征融合模块，也可以视为一种利用辅助信息(Decoder)特征以增强Encoder特征的增强模块

---

模块代码

import torch
import torch.nn as nn
import torch.nn.functional as F


class GAG(nn.Module):
    def __init__(self, F_g, F_l, F_int, kernel_size=1, groups=1):
        super(GAG,self).__init__()
        if kernel_size == 1:
            groups = 1
        self.W_g = nn.Sequential(
            nn.Conv2d(F_g, F_int, kernel_size=kernel_size,stride=1,padding=kernel_size//2,groups=groups, bias=True),
            nn.BatchNorm2d(F_int)
        )
        self.W_x = nn.Sequential(
            nn.Conv2d(F_l, F_int, kernel_size=kernel_size,stride=1,padding=kernel_size//2,groups=groups, bias=True),
            nn.BatchNorm2d(F_int)
        )
        self.psi = nn.Sequential(
            nn.Conv2d(F_int, 1, kernel_size=1,stride=1,padding=0,bias=True),
            nn.BatchNorm2d(1),
            nn.Sigmoid()
        )
        self.activation = nn.ReLU(inplace=True)

        
    def forward(self,g,x):
        g1 = self.W_g(g)
        x1 = self.W_x(x)
        psi = self.activation(g1+x1)
        psi = self.psi(psi)
        return x*psi
    

if __name__ == '__main__':
    x1 = torch.randn([1, 64, 44, 44])
    x2 = torch.randn([1, 64, 44, 44])
    gag = GAG(F_g=64, F_l=64, F_int=64//2, kernel_size=3, groups=64//2)
    out = gag(x1, x2)
    print(out.shape)  # [1, 64, 44, 44]

---