Computes Dice Loss according to https://arxiv.org/abs/1606.04797. For multi-class segmentation `weight` parameter can be used to assign different

Code for reading the infant brain MICCAIBraTS2018 challenge

Code for reading the COVID Segmentation dataset Segmentation Task 1: Learning with limited annotations This task is based on the CO

Code for reading the COVIDxDataset

Code for reading Gleason 2019 MICCAI Challenge

Code for reading the infant brain MRI dataset of ISEG 2017 challenge

(conv => BN => ReLU) * 2

Code for reading the infant brain MRI dataset of ISEG 2017 challenge

Linear combination of BCE and Dice losses3D

Computes Generalized Dice Loss (GDL) as described in https://arxiv.org/pdf/1707.03237.pdf.

Code for reading the IXI brain MRI dataset This loader is implemented for cross-dataset testing

Code for reading the infant brain MICCAIBraTS2018 challenge

Code for reading the infant brain MICCAIBraTS2018 challenge

WeightedCrossEntropyLoss (WCE) as described in https://arxiv.org/pdf/1707.03237.pdf

to keep the spatial dims o=i, this formula is applied o = [i + 2*p - k - (k-1)*(d-1)]/s + 1

Implementation of contrastive loss defined in https://arxiv.org/pdf/1708.02551.pdf 'Semantic Instance Segmentation with a Discriminative Loss

Implementation based on https://arxiv.org/abs/1708.00573 Trainable params: 1,783,408 (roughly 1.8 mentioned in the paper)

2-stream and 3-stream implementation with early fusion dual-single-densenet OR Disentangled modalities with early fusion in the paper

Densely Connected Convolutional Networks" <https://arxiv.org/pdf/1608.06993.pdf>` Based on the implementation of https://github.com/tbuikr/3D-Skip

Trainer class

(segmentation)we transfer encoder part from Med3D as the feature extraction part and then segmented lung in whole body followed by three gro

Implementations based on the Unet3D paper: https://arxiv.org/abs/1606.06650

Implementations based on the Vnet paper: https://arxiv.org/abs/1606.04797

A lighter version of Vnet that skips down_tr256 and up_tr256 in oreder to reduce time and space complexity

Constructs a series of dense-layers based on in and out kernels list

For transpose conv o = output, p = padding, k = kernel_size, s = stride, d = dilation o = (i -1)*s - 2*p + k + output_padding = (i-1)*2 +

r""" BaseModel with basic functionalities for checkpointing and restoration.

Base class for all trainers

Composes several transforms together.

choose a random tranform from list an apply transforms: tranforms to apply p: probability

Loss wrapper which removes additional target channel

TODO fix transpose conv to double spatial dim

Base class for different implementations of Dice loss.

Loss wrapper which prevents the gradient of the loss to be computed where target is equal to `ignore_index`.