Published October 26, 2020
| Version 1.0
Software
Open
nnU-Net Singularity Container for SAX cMRI Segmentation Trained on Data from the M&Ms Challenge 2020
- 1. Division of Medical Image Computing, German Cancer Research Center (DKFZ); Heidelberg University, Medical Faculty Heidelberg, Heidelberg, Germany)
- 2. Division of Medical Image Computing, German Cancer Research Center (DKFZ)
Description
Final submission Singularity container for the winning contribution of the Multi-Centre, Multi-Vendor & Multi-Disease Cardiac Image Segmentation Challenge (M&Ms).
The model inside container was trained with the nnU-Net framework. In order to be able to run this container you will need a GPU comparable to TITAN Xp (12 GB VRAM) or larger.
Follow these steps to perform fully automatic segmentation on short axis (SAX) cMRI data:
- install Singularity: Follow instructions by sylabs.
- prepare your data as 3D .nii.gz files
- put your data in a folder, that will serve as your input folder (<input_folder>)
- to make your folder structure conform with the folder structure of the challenge place a subfolder called "mnms" inside <input_folder> as shown below
<input_folder>
|- subfolder
|-- image01.nii.gz
|-- image02.nii.gz
|-- ...
Then run the container
singularity run --nv <path_to_container_folder>/MIC_DKFZ_mnms_final_submission.sif <input_folder> <output_folder>
The nnU-Net inside the container will automatically ensemble five 2D and five 3D pretrained U-Nets and will save the final prediction in your defined <output_folder>.
Files
Files
(4.1 GB)
Name | Size | Download all |
---|---|---|
md5:0acbb80eaac1399e5c71ee99bcef8c03
|
4.1 GB | Download |
Additional details
References
- Fabian Isensee, Paul F. Jäger, Simon A. A. Kohl, Jens Petersen, Klaus H. Maier-Hein "Automated Design of Deep Learning Methods for Biomedical Image Segmentation" arXiv preprint arXiv:1904.08128 (2020).
- Campello, Víctor M. et al.: Multi-Centre, Multi-Vendor & Multi-Disease Cardiac Image Segmentation. In preparation.