Table 3.

Detailed configuration of the spatial attention module

Layers	Output shape
$♡$ Input	[1024, 376]
$♠$ MaxPool(1)	[1, 376]
$♣$ AvgPool(1)	[1, 376]
Concat($♠$, $♣$)	[2, 376]
$♢$ Conv1D(1, 7, 1) $+$ Sigmoid	[1, 376]
Multiply($♡$, $♢$)	[1, 376]

Note: $♠$ and $♣$ are in parallel. ‘Concat’ and ‘Multiply’ denote concatenation and element-wise multiplication, respectively