SPINEPS – Automatic Whole Spine Segmentation of T2w MR images using a Two-Phase Approach to Multi-class Semantic and Instance Segmentation.

This is a segmentation pipeline to automatically, and robustly, segment the whole spine in T2w sagittal images.

Citation

If you are using SPINEPS, please cite the following:

Hendrik Möller, Robert Graf, Joachim Schmitt, Benjamin Keinert, Matan Atad, Anjany
Sekuboyina, Felix Streckenbach, Hanna Schon, Florian Kofler, Thomas Kroencke, Ste-
fanie Bette, Stefan Willich, Thomas Keil, Thoralf Niendorf, Tobias Pischon, Beate Ende-
mann, Bjoern Menze, Daniel Rueckert, and Jan S. Kirschke. Spineps – automatic whole
spine segmentation of t2-weighted mr images using a two-phase approach to multi-class
semantic and instance segmentation. arXiv preprint arXiv:2402.16368, 2024.

ArXiv link: https://arxiv.org/abs/2402.16368

BibTeX citation:

@article{moeller2024,
      title={SPINEPS -- Automatic Whole Spine Segmentation of T2-weighted MR images using a Two-Phase Approach to Multi-class Semantic and Instance Segmentation},
      author={Hendrik Möller and Robert Graf and Joachim Schmitt and Benjamin Keinert and Matan Atad and Anjany Sekuboyina and Felix Streckenbach and Hanna Schön and Florian Kofler and Thomas Kroencke and Stefanie Bette and Stefan Willich and Thomas Keil and Thoralf Niendorf and Tobias Pischon and Beate Endemann and Bjoern Menze and Daniel Rueckert and Jan S. Kirschke},
    journal={arXiv preprint arXiv:2402.16368},
    year={2024},
    eprint={2402.16368},
    archivePrefix={arXiv},
    primaryClass={eess.IV},
}

Installation (Ubuntu)

This installation assumes you know your way around conda and virtual environments.

Setup Venv

The order of the following instructions is important!

1. Use Conda or Pip to create a venv for python 3.11, we are using conda for this example:

conda create --name spineps python=3.11
conda activate spineps
conda install pip

2. Go to https://pytorch.org/get-started/locally/ and install a correct pytorch version for your machine in your venv
3. Confirm that your pytorch package is working! Try calling these commands:

nvidia-smi

This should show your GPU and it's usage.

python -c "import torch; print(torch.cuda.is_available())"

This should throw no errors and return True

Setup this package

You have to install the package to use it, even if you just want to locally use the code.

1. cd into the spineps folder and install it by running pip install -e . or using the pyproject.toml inside of the project folder.
2. Download the model weights from https://syncandshare.lrz.de/getlink/fi16bYYmqpwPQZRGd1M4G6/
3. Extract the downloaded modelweights folders into a folder of your choice (the "spineps/spineps/models" folders will be used as default), from now on referred to as your models folder. This specified folder should have the following structure:

<models_folder>
├── <model_name 1>
    ├── inference_config.json
    ├── <other model-specific files and folders>
├── <model_name 2>
    ├── inference_config.json
    ├── <other model-specific files and folders>
...

3. You need to specify this models folder as argument when running. If you want to set it permanently, set the according environment variable in your .bashrc or .zshrc (whatever you are using).

export SPINEPS_SEGMENTOR_MODELS=<PATH-to-your-folder>

You can also execute the above line whenever you run this segmentation pipeline.

To check that you set the environment variable correctly, call:

echo ${SPINEPS_SEGMENTOR_MODELS}

For Windows, this might help: https://phoenixnap.com/kb/windows-set-environment-variable

If you don't set the environment variable, the pipeline will look into spineps/spineps/models/ by default.

Usage

Installed as package:

1. Activate your venv
2. Run spineps -h to see the arguments

Installed as local clone:

1. Activate your venv
2. Run python entrypoint.py -h to see the arguments.
3. For example, for a sample, run python entrypoint.py sample -i <path-to-nifty> -model_semantic <model_name> -model_instance <model_name> (replacing <model_name> with the name of the model you want to use)

Issues

• import issues: try installing via the requirements again, somethings it doesn't install everything
• pytorch / cuda issues: good luck! :3

SPINEPS Capabilities

The pipeline can process either:

• Single Nifty (.nii.gz) files
• Whole Datasets

Single nifty

spineps sample <args>:

Processes a single nifty file, will create a derivatves folder next to the nifty, and write all outputs into that folder

argument	explanation
-i	Absolute path to the single nifty file (.nii.gz) to be processed
-model_semantic , -ms	The model used for the semantic segmentation
-model_instance , -mv	The model used for the vertebra instance segmentation
-der_name , -dn	Name of the derivatives folder (default: derivatives_seg)
-save_debug, -sd	Saves a lot of debug data and intermediate results in a separate debug-labeled folder (default: False)
-save_unc_img, -sui	Saves a uncertainty image from the subreg prediction (default: False)
-override_semantic, -os	Will override existing seg-spine files (default: False)
-override_instance, -ov	Will override existing seg-vert files (default: False)
-override_ctd, -oc	Will override existing centroid files (default: False)
-verbose, -v	Prints much more stuff, may fully clutter your terminal (default: False)

There are a lot more arguments, run spineps sample -h to see them.

Dataset

spineps dataset <args>:

Processes all "suitable" niftys it finds in the specified dataset folder.

A dataset folder must have the following structure:

dataset-folder
├── <rawdata>
    ├── subfolders (optionally, any number of them)
        ├── One or multiple target files
    ├── One or multiple target files
├── <derivatives>
    ├── The results are saved/loaded here

A target file in a dataset must look like the following:

sub-<subjectid>_*_T2w.nii.gz

where * depicts any number of key-value pairs of characters. Some examples are:

sub-0001_T2w.nii.gz
sub-awesomedataset_sequ-HWS_part-inphase_T2w.nii.gz

Anything that follows the BIDS-nomenclature is also supported (see https://bids-specification.readthedocs.io/en/stable/) Meaning you can have some key-value pairs (like sub-<id>) in the name. Those key-value pairs are always separated by _ and combined with - (see second example above). Those will be used in creating the filename of the created segmentations.

To that end, we are using TPTBox (see https://github.com/Hendrik-code/TPTBox)

It supports the same arguments as in sample mode (see table above), and additionally:

argument	explanation
-raw_name, -rn	Sets the name of the rawdata folder of the dataset (default: "rawdata")
-ignore_bids_filter, -ibf	If true, will search the BIDS dataset without the strict filters. Use with care! (default: False)
-ignore_model_compatibility, -imc	If true, will not stop the pipeline to use the given models on unfitting input modalities (default: False)
-save_log, -sl	If true, saves the log into a separate folder in the dataset directory (default: False)
-save_snaps_folder, -ssf	If true, additionally saves the snapshots in a separate folder in the dataset directory (default: False)

For a full list of arguments, call spineps dataset -h

Segmentation

The pipeline segments in multiple steps:

1. Semantically segments 14 spinal structures (9 regions for vertebrae, Spinal Cord, Spinal Canal, Intervertebral Discs, Endplate, Sacrum)
2. From the vertebra regions, segment the different vertebrae as instance mask
3. Save the first as seg-spine mask, the second as seg-vert mask
4. It can save an uncertainty image for the semantic segmentation
5. From the two segmentations, calculates centroids for each vertebrae center point, endplate, and IVD and saves that into a .json
6. From the centroid and the segmentations, makes a snapshot showcasing the result as a .png

Labels:

In the subregion segmentation:

Label	Structure
41	Arcus_Vertebrae
42	Spinosus_Process
43	Costal_Process_Left
44	Costal_Process_Right
45	Superior_Articular_Left
46	Superior_Articular_Right
47	Inferior_Articular_Left
48	Inferior_Articular_Right
49	Vertebra_Corpus_border
60	Spinal_Cord
61	Spinal_Canal
62	Endplate
100	Vertebra_Disc
26	Sacrum

In the vertebra instance segmentation mask, each label X in [1, 25] are the unique vertebrae, while 100+X are their corresponding IVD and 200+X their endplates.

Using the Code

If you want to call the code snippets yourself, start by initializing your models using seg_model.get_segmentation_model() giving it the absolute path to your model folder.

Depending on whether you want to process a single sample or a whole dataset, go into seg_run.py and run either process_img_nii() or process_dataset().

If you want to perform even more detailed changes or code injections, see process_img_nii() as inspiration on how the underlaying functions work and behave. Treat with care!

Authorship

This pipeline was created by Hendrik Möller, M.Sc. (he/him)
PhD Researcher at Department for Interventional and Diagnostic Neuroradiology

Developed within an ERC Grant at
University Hospital rechts der Isar at Technical University of Munich
Ismaninger Street 22, 81675 Munich

https://deep-spine.de/
https://aim-lab.io/author/hendrik-moller/

License

Copyright 2023 Hendrik Möller

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

http://www.apache.org/licenses/LICENSE-2.0

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.