sksurgerytorch.models.high_res_stereo_model module

Definition of the HSMNet model structure, and various helper functions.

class sksurgerytorch.models.high_res_stereo_model.HSMNet_model(maxdisp, clean, device, level=1)

Bases: torch.nn.modules.module.Module

feature_vol(refimg_fea, targetimg_fea, maxdisp, leftview=True)

diff feature volume

forward(left, right)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class sksurgerytorch.models.high_res_stereo_model.conv2DBatchNorm(in_channels, n_filters, k_size, stride, padding, bias=True, dilation=1, with_bn=True)

Bases: torch.nn.modules.module.Module

forward(inputs)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class sksurgerytorch.models.high_res_stereo_model.conv2DBatchNormRelu(in_channels, n_filters, k_size, stride, padding, bias=True, dilation=1, with_bn=True)

Bases: torch.nn.modules.module.Module

forward(inputs)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class sksurgerytorch.models.high_res_stereo_model.decoderBlock(nconvs, inchannelF, channelF, stride=(1, 1, 1), up=False, nstride=1, pool=False)

Bases: torch.nn.modules.module.Module

forward(fvl)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class sksurgerytorch.models.high_res_stereo_model.disparityregression(maxdisp, divisor)

Bases: torch.nn.modules.module.Module

forward(x, ifent=False)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class sksurgerytorch.models.high_res_stereo_model.projfeat3d(in_planes, out_planes, stride)

Bases: torch.nn.modules.module.Module

Turn 3d projection into 2d projection

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class sksurgerytorch.models.high_res_stereo_model.pyramidPooling(in_channels, pool_sizes, model_name='pspnet', fusion_mode='cat', with_bn=True)

Bases: torch.nn.modules.module.Module

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class sksurgerytorch.models.high_res_stereo_model.residualBlock(in_channels, n_filters, stride=1, downsample=None, dilation=1)

Bases: torch.nn.modules.module.Module

expansion = 1

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

sksurgerytorch.models.high_res_stereo_model.sepConv3d(in_planes, out_planes, kernel_size, stride, pad, bias=False)

class sksurgerytorch.models.high_res_stereo_model.sepConv3dBlock(in_planes, out_planes, stride=(1, 1, 1))

Bases: torch.nn.modules.module.Module

Separable 3d convolution block as 2 separable convolutions and a projection layer

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class sksurgerytorch.models.high_res_stereo_model.unet

Bases: torch.nn.modules.module.Module

forward(x)

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.