fxxJuses/DeepLearning_UCAS2022_registration

DeepLearning_UCAS2022_registration

This repository is used to manage our team's code and document sharing. I hope everyone can gain a lot from this cooperation.

Dataset

You can download the dataset in MICCAI chanllenge ,url: https://learn2reg.grand-challenge.org/Datasets/

or You can download the dataset in Baidu cloud: 链接：https://pan.baidu.com/s/1GN2_7KHCeZEvX5ECdQ4A3g 提取码：wsgk

The validation and test dataset have been published in Release

Literature sharing

We hope that we can have some insights and record them in the process of research. It is recommended to use the markdown .

Requirement installation

conda install pytorch torchvision torchaudio cudatoolkit=11.3 -c pytorch
conda install batchgenerators
conda install natsort
conda install pystrum
conda install nibabel
conda install ml_collections
conda install timm

Models

If you want to reproduce our work, you can do it with the following command !

VoxelMorph

• Train

python ./script/VoxelMoprh/train.py

• Test

python ./script/VoxelMoprh/infer.py

• file description

file/folder	function
data	Generate dataset inputs
Model	The overall framework of the VoxelMorph model
infer.py	test model
losses.py	Loss function
seg35_labels.txt	Segmentation map label description, used to finally calculate the dsc for each class
train.py	train model
utils.py	Basic functions and classes that compute some metrics and implement some basic transformations

CycleMorph

• Train

python ./script/CycleMoprh/train_CycleMorph.py

• Test

python ./script/CycleMoprh/infer.py

• file description

file/folder	function
data	Generate dataset inputs
Model	The overall framework of the CycleMoprh model
infer.py	test model
losses.py	Loss function
seg35_labels.txt	Segmentation map label description, used to finally calculate the dsc for each class
train_CycleMorph.py	train model
utils.py	Basic functions and classes that compute some metrics and implement some basic transformations

ConsMorph

• Train

python ./script/ConsMoprh/train_Consmorph.py

• Test

python ./script/ConsMoprh/infer.py

• file description

file/folder	function
data	Generate dataset inputs
Model	The overall framework of the ConsMoprh model
infer.py	test model
losses.py	Loss function
seg35_labels.txt	Segmentation map label description, used to finally calculate the dsc for each class
train_CycleMorph.py	train model
utils.py	Basic functions and classes that compute some metrics and implement some basic transformations

TransMorph

• Train

  python ./script/TransMorph/train_TransMorph.py

• Test

  python ./script/TransMorph/infer_TransMorph.py

• 代码文件说明

  file/folder	作用
  data	Generate dataset inputs
  models	The overall framework of the TransMorph model
  convert_nii_to_pkl.py	Convert nii.gz file to pkl
  infer_TransMorph.py	test model
  losses.py	Loss function
  seg35_labels.txt	Segmentation map label description, used to finally calculate the dsc for each class
  train_TransMorph.py	train model
  utils.py	Basic functions and classes that compute some metrics and implement some basic transformations

nnFormer

• 训练模型

  python ./script/nnFormer/train_TransMorph.py

• 测试模型

  python ./script/nnFormer/infer_TransMorph.py

• 代码文件说明

  文件/文件夹	作用
  data	Generate dataset inputs
  models	The overall framework of the nnFormer model
  img0438.nii.gz	The data of the subject numbered 0438 in the test set is used to obtain affine in the test phase
  infer_nnFormer.py	test model
  losses.py	Loss function
  seg35_labels.txt	Segmentation map label description, used to finally calculate the dsc for each class
  train_nnFormer.py	train model
  utils.py	计算一些指标和实现一些基本变换的基本函数和类
  val_nnFormer.py	Evaluate the model

SyN

• 测试模型

  python ./script/SyN/infer_SyN.py

• 代码文件说明

  文件/文件夹	作用
  data	Generate dataset inputs
  infer_SyN.py	test model
  seg35_labels.txt	Segmentation map label description, used to finally calculate the dsc for each class
  utils.py	计算一些指标和实现一些基本变换的基本函数和类

Reference:

TransMorph