Table 1.

Some of the symbols in this chapter.

Symbol	Paraphrase
H	3 × 3 convolution block with a LeakyRelu activation
C	Convolution block with 3 × 3 core and stride 2
T	Transpose convolution layer with 3 × 3 core and stride 2
F	Residual block with 2 H operations and a skip connection
SE	Squeeze and excitation block
s	Representation of h(high) in high-res routes and l(low)in low-res routes
w	The scaling factor
D n s	The input of the nth H structure
M n s	The output map of the nth H structure
$\widehat{D_m^a}$ , $\widehat{D_n^b}$	The output of the RFN subnetwork
S l	Shape stream feature map of the lth layer
A SC	The output of spatial and channel attention block
A AG	The output of the attention gate block
P	The output of the previous decoder block