liom_toolkit.segmentation.vseg.model module

class liom_toolkit.segmentation.vseg.model.ConvBlock(in_c, out_c)

Bases: Module

Convolutional block for the U-Net architecture

forward(inputs: Tensor) → Tensor

Forward pass of the convolutional block

Parameters:

inputs (torch.Tensor) – The input tensor

Returns:

The output tensor

Return type:

torch.Tensor

class liom_toolkit.segmentation.vseg.model.DecoderBlock(in_c, out_c)

Bases: Module

Decoder block for the U-Net architecture

forward(inputs: Tensor, skip: Tensor) → Tensor

Forward pass of the decoder block

Parameters:

• inputs (torch.Tensor) – The input tensor

• skip (torch.Tensor) – The skip tensor

Returns:

The output tensor

Return type:

torch.Tensor

class liom_toolkit.segmentation.vseg.model.EncoderBlock(in_c, out_c)

Bases: Module

Encoder block for the U-Net architecture

forward(inputs: ~torch.Tensor) -> (<class 'torch.Tensor'>, <class 'torch.Tensor'>)

Forward pass of the encoder block

Parameters:

inputs (torch.Tensor) – The input tensor

Returns:

The output tensor

Return type:

torch.Tensor

class liom_toolkit.segmentation.vseg.model.VsegModel(pretrained: bool = False, device: device = device(type='cpu'))

Bases: Module

U-Net model for vessel segmentation

b

Decoder

d4

Classifier

e4

Bottleneck

forward(inputs: Tensor) → Tensor

Forward pass of the U-Net model

Parameters:

inputs (torch.Tensor) – The input tensor

Returns:

The output tensor

Return type:

torch.Tensor