def __init__(self, in_channels, out_channels, num_classes):
super(RCCAModule, self).__init__()
inter_channels = in_channels // 4
self.conva = nn.Sequential(
nn.Conv2d(in_channels, inter_channels, 3, padding=1, bias=False),
InPlaceABNSync(inter_channels))
self.cca = CrissCrossAttention(inter_channels)
self.convb = nn.Sequential(
nn.Conv2d(inter_channels, inter_channels, 3, padding=1,
bias=False), InPlaceABNSync(inter_channels))
self.bottleneck = nn.Sequential(
nn.Conv2d(in_channels + inter_channels,
out_channels,
kernel_size=3,
padding=1,
dilation=1,
bias=False), InPlaceABNSync(out_channels),
nn.Dropout2d(0.1),
nn.Conv2d(512,
num_classes,
kernel_size=1,
stride=1,
padding=0,
bias=True))
def __init__(self, in_channels, out_channels):
super(RCCAModule, self).__init__()
inter_channels = in_channels // 4
self.conva = nn.Sequential(nn.Conv2d(in_channels, inter_channels, 3, padding=1, bias=False),
nn.BatchNorm2d(inter_channels))
self.cca = CrissCrossAttention(inter_channels)
self.convb = nn.Sequential(nn.Conv2d(inter_channels, out_channels, 3, padding=1, bias=False),
nn.BatchNorm2d(out_channels))
def __init__(self, num_classes, n_bands, chanel):
super(SSCDNonLModel, self).__init__()
#self.num_Node=num_Node
self.bands = n_bands
chanel = chanel
kernel = 5
CCChannel = 25
self.b1 = nn.BatchNorm2d(self.bands)
self.con1 = nn.Conv2d(self.bands, chanel, 1, padding=0, bias=True)
self.s1 = nn.Sigmoid()
self.cond1 = nn.Conv2d(chanel,
chanel,
kernel,
padding=2,
groups=chanel,
bias=True)
self.sd1 = nn.Sigmoid()
self.b2 = nn.BatchNorm2d(self.bands + chanel)
#self.nlcon1=NonLocalBlock(300, 300, True)
#self.gcn1=GCNtrans(300,1000)
#self.bcat=nn.BatchNorm2d(300+300)
self.con2 = nn.Conv2d(self.bands + chanel,
chanel,
1,
padding=0,
bias=True)
self.s2 = nn.Sigmoid()
self.cond2 = nn.Conv2d(chanel,
CCChannel,
kernel,
padding=2,
groups=25,
bias=True)
self.sd2 = nn.Sigmoid()
self.b4 = nn.BatchNorm2d(CCChannel)
self.nlcon2 = CrissCrossAttention(CCChannel)
self.nlcon3 = CrissCrossAttention(CCChannel)
self.bcat = nn.BatchNorm2d(CCChannel + CCChannel)
self.con4 = nn.Conv2d(CCChannel + CCChannel,
chanel,
1,
padding=0,
bias=True)
self.s4 = nn.Sigmoid()
self.cond4 = nn.Conv2d(chanel,
chanel,
kernel,
padding=2,
groups=chanel,
bias=True)
self.sd4 = nn.Sigmoid()
self.b5 = nn.BatchNorm2d(CCChannel + chanel)
self.con5 = nn.Conv2d(CCChannel + chanel,
chanel,
1,
padding=0,
bias=True)
self.s5 = nn.Sigmoid()
self.cond5 = nn.Conv2d(chanel,
chanel,
kernel,
padding=2,
groups=chanel,
bias=True)
self.sd5 = nn.Sigmoid()
#self.b6=nn.BatchNorm2d(300+300)
self.con6 = nn.Conv2d(chanel + CCChannel,
num_classes + 1,
1,
padding=0,
bias=True)