AttentionModule

tiatoolbox . models . architecture . utils . AttentionModule

class AttentionModule(name, in_channels, reduction=16)

Attention module to apply attention mechanism on feature maps.

Initialize the Attention module.

Parameters:

• name (str | None) – Name of the attention mechanism. Only “scse” is implemented. If None, identity is used.

• in_channels (int) – Number of input channels.

• reduction (int) – Reduction ratio for channel attention.

Methods

forward

Forward pass of the Attention module.

Attributes

training

forward(x)

Forward pass of the Attention module.

Parameters:

• x (torch.Tensor) – Input feature map of shape (N, C, H, W).

• self (AttentionModule)

Returns:

Output feature map after applying attention.

Return type:

torch.Tensor