File size: 5,631 Bytes
6bd0db7 9363120 6bd0db7 9363120 6bd0db7 9363120 6bd0db7 9363120 6bd0db7 9363120 6bd0db7 9363120 6bd0db7 9363120 6bd0db7 9363120 | 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 | import torch
import torch.nn as nn
import torch.nn.functional as F
from torchvision import models
class DoubleConv(nn.Module):
def __init__(self, in_channels, out_channels, dropout=0.0):
super(DoubleConv, self).__init__()
layers = [
nn.Conv2d(in_channels, out_channels, kernel_size=3, padding=1),
nn.BatchNorm2d(out_channels),
nn.ReLU(inplace=True),
nn.Conv2d(out_channels, out_channels, kernel_size=3, padding=1),
nn.BatchNorm2d(out_channels),
nn.ReLU(inplace=True),
]
if dropout > 0:
layers.append(nn.Dropout2d(p=dropout))
self.net = nn.Sequential(*layers)
def forward(self, x):
return self.net(x)
def crop_to_match(enc_feat, dec_feat):
"""Center-crop encoder feature map to match size of decoder feature map."""
_, _, H, W = dec_feat.shape
enc_H, enc_W = enc_feat.shape[2], enc_feat.shape[3]
crop_top = (enc_H - H) // 2
crop_left = (enc_W - W) // 2
return enc_feat[:, :, crop_top:crop_top+H, crop_left:crop_left+W]
class UNet(nn.Module):
def __init__(self, in_channels=1, out_channels=1, dropout=0.1):
super(UNet, self).__init__()
# Encoder
self.enc1 = DoubleConv(in_channels, 64, dropout=dropout)
self.enc2 = DoubleConv(64, 128, dropout=dropout)
self.enc3 = DoubleConv(128, 256, dropout=dropout)
self.enc4 = DoubleConv(256, 512, dropout=dropout)
self.pool = nn.MaxPool2d(2)
self.bottleneck = DoubleConv(512, 1024, dropout=dropout)
# Decoder
self.up4 = nn.ConvTranspose2d(1024, 512, kernel_size=2, stride=2)
self.dec4 = DoubleConv(1024, 512, dropout=dropout)
self.up3 = nn.ConvTranspose2d(512, 256, kernel_size=2, stride=2)
self.dec3 = DoubleConv(512, 256, dropout=dropout)
self.up2 = nn.ConvTranspose2d(256, 128, kernel_size=2, stride=2)
self.dec2 = DoubleConv(256, 128, dropout=dropout)
self.up1 = nn.ConvTranspose2d(128, 64, kernel_size=2, stride=2)
self.dec1 = DoubleConv(128, 64, dropout=dropout)
self.final = nn.Conv2d(64, out_channels, kernel_size=1)
def forward(self, x):
input_size = x.shape[2:] # (H, W)
# Encoder
e1 = self.enc1(x)
e2 = self.enc2(self.pool(e1))
e3 = self.enc3(self.pool(e2))
e4 = self.enc4(self.pool(e3))
# Bottleneck
b = self.bottleneck(self.pool(e4))
# Decoder with cropping
d4 = self.up4(b)
e4_cropped = crop_to_match(e4, d4)
d4 = self.dec4(torch.cat([d4, e4_cropped], dim=1))
d3 = self.up3(d4)
e3_cropped = crop_to_match(e3, d3)
d3 = self.dec3(torch.cat([d3, e3_cropped], dim=1))
d2 = self.up2(d3)
e2_cropped = crop_to_match(e2, d2)
d2 = self.dec2(torch.cat([d2, e2_cropped], dim=1))
d1 = self.up1(d2)
e1_cropped = crop_to_match(e1, d1)
d1 = self.dec1(torch.cat([d1, e1_cropped], dim=1))
out = self.final(d1)
# Resize output back to input size (200x200)
out = F.interpolate(out, size=input_size, mode="bilinear", align_corners=False)
return out
# # =========================================================
# # 1. U-Net
# # =========================================================
# class UNet(nn.Module):
# def __init__(self, in_channels=1, out_channels=1):
# super().__init__()
# def CBR(in_c, out_c):
# return nn.Sequential(
# nn.Conv2d(in_c, out_c, 3, padding=1),
# nn.BatchNorm2d(out_c),
# nn.ReLU(inplace=True)
# )
# self.enc1 = nn.Sequential(CBR(in_channels, 64), CBR(64, 64))
# self.enc2 = nn.Sequential(CBR(64, 128), CBR(128, 128))
# self.enc3 = nn.Sequential(CBR(128, 256), CBR(256, 256))
# self.enc4 = nn.Sequential(CBR(256, 512), CBR(512, 512))
# self.pool = nn.MaxPool2d(2, 2)
# self.center = nn.Sequential(CBR(512, 1024), CBR(1024, 512))
# self.up4 = nn.ConvTranspose2d(512, 512, 2, stride=2)
# self.dec4 = nn.Sequential(CBR(1024, 512), CBR(512, 256))
# self.up3 = nn.ConvTranspose2d(256, 256, 2, stride=2)
# self.dec3 = nn.Sequential(CBR(512, 256), CBR(256, 128))
# self.up2 = nn.ConvTranspose2d(128, 128, 2, stride=2)
# self.dec2 = nn.Sequential(CBR(256, 128), CBR(128, 64))
# self.up1 = nn.ConvTranspose2d(64, 64, 2, stride=2)
# self.dec1 = nn.Sequential(CBR(128, 64), nn.Conv2d(64, out_channels, 1))
# def crop_to_match(self, x, target):
# _, _, h, w = target.size()
# return x[:, :, :h, :w]
# def forward(self, x):
# e1 = self.enc1(x)
# e2 = self.enc2(self.pool(e1))
# e3 = self.enc3(self.pool(e2))
# e4 = self.enc4(self.pool(e3))
# c = self.center(self.pool(e4))
# d4 = self.up4(c)
# e4_c = self.crop_to_match(e4, d4)
# d4 = torch.cat([d4, e4_c], dim=1)
# d4 = self.dec4(d4)
# d3 = self.up3(d4)
# e3_c = self.crop_to_match(e3, d3)
# d3 = torch.cat([d3, e3_c], dim=1)
# d3 = self.dec3(d3)
# d2 = self.up2(d3)
# e2_c = self.crop_to_match(e2, d2)
# d2 = torch.cat([d2, e2_c], dim=1)
# d2 = self.dec2(d2)
# d1 = self.up1(d2)
# e1_c = self.crop_to_match(e1, d1)
# d1 = torch.cat([d1, e1_c], dim=1)
# out = self.dec1(d1)
# out = F.interpolate(out, size=x.shape[2:], mode="bilinear", align_corners=False)
# return out
|