Multivariate mixture model in PyTorch

This is a course project in Medical Image Analysis on “Multivariate mixture model for myocardial segmentation combining multi-source images”. The algorithm is re-implemented in PyTorch. The lecture notes for illustration and presentation is also included.

Getting Started

Project structure

The project contains PyTorch implementation of the algorithm from Multivariate mixture model on 2D multi-sequence cardiac MR images. The data can be downloaded from MS-CMRSeg-2019 challenge. The project structure is as follows:

MvMM-Demo
|-- src
|   |-- AffineGrid.py                # convert affine matrix to resampling grid
|   |-- LocalDisplacementEnergy.py   # displacement regularization, bending energy
|   |-- MvMMVEM.py                   # model construction and EM algorithm
|   |-- MvMMVEMDemo.py               # Demo: image loading, preprocessing, model optimization and result visualization
|   |-- SpatialTransformer.py        # spatial transformation module
|   |-- image_utils.py               # functions for image loading and preprocessing
|   |-- metrics.py                   # metrics computation
|   |-- utils.py                     # utility functions

Usage

Combined segmentation from a set of images is achieved by:

python MvMMVEMDemo.py 
--data_path #YOUR OWN DATA PATH#           # data path to load images
--image_names #YOUR OWN IMAGE NAMES#       # image names
--atlas_name #YOUR OWN ARLAS NAME#         # atlas name
--label_intensities 0 255                  # label intensity values
--vol_shape 256 256                        # image size
--num_subjects 3                           # number of subjects
--num_classes 2                            # number of classes
--num_subtypes 2 2                         # number of subtypes
--transform rigid                          # transformation type
--training_iters 1000                      # training iterations
--EM_steps 3                               # EM update steps
--bending_energy 1                         # bending energy coefficient

Citation

If you found the project useful, please cite our papers as below:

@article{Zhuang2019MvMM,
  title={Multivariate mixture model for myocardial segmentation combining multi-source images},
  author={Zhuang, Xiahai},
  journal={IEEE transactions on pattern analysis and machine intelligence},
  volume={41},
  number={12},
  pages={2933--2946},
  year={2019},
  publisher={IEEE}
}

@inproceedings{Luo2020MvMM-RegNet,
  title={MvMM-RegNet: A new image registration framework based on multivariate mixture model and neural network estimation},
  author={Luo, Xinzhe and Zhuang, Xiahai},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  pages={149--159},
  year={2020},
  organization={Springer}
}

Contact

For any questions or problems please open an issue on GitHub.