Stand-alone project for segmentation validation and metrics computation

Designed for brain tumor segmentation purposes. This software was released together with the article Preoperative brain tumor imaging: models and software for segmentation and standardized reporting.

1. Folder and data structures and naming conventions

Assuming in the following example that the data are stored based on their origin, but anything should work. The folders named index0 and index1 could be renamed to Amsterdam and StOlavs for instance.

1.1 Original data folder structure

The main data directory containing the original MRI images and corresponding manual annotations is expected to resemble:

└── path/to/data/root/
    └── index0/
        ├── AMS0/
        │   ├── volumes/
        │   │   └── AMS0_T1.nii.gz
        │   └── segmentations/
        │       └── AMS0_T1_label_tumor.nii.gz
        ├── AMS25/
        └── AMS50/
    └── index1/
        ├── STO25/
        └── STO50/

1.2 Inference results folder structure

The inference results should be grouped inside what will become the validation folder, resembling the following structure (here for Study1). The outer-most sub-folder naming convention inside predictions are the fold numbers.

└── path/to/validation/study/
    └── predictions/
        ├── 0/
        │   ├── index0_AMS0/
        │   |   └── AMS0_T1-predictions.nii.gz  
        │   ├── index1_STO25/ 
        │   |   └── STO25_T1-predictions.nii.gz  
        └── 0/
            ├── index0_AMS50/
            │   └── AMS50_T1-predictions.nii.gz  
            └── index1_STO50/ 
                └── STO50_T1-predictions.nii.gz

1.3 Folds file

The file with patients' distribution within each fold used for training should list the content of the validation and test sets iteratively.
The file should be called cross_validation_folds.txt and placed in the validation study folder side-by-side with the predictions sub-folder.
An example of its content is given below:

index0_AMS1000_T1_sample index1_STO250_T1_sample\n
index0_AMS0_T1_sample index1_STO25_T1_sample\n
index0_AMS0_T1_sample index1_STO25_T1_sample\n
index0_AMS25_T1_sample index1_STO50_T1_sample\n

2. Installation

Create a virtual environment using at least Python 3.7.0, and install all dependencies from the requirements.txt file.

cd /path/to/validation_metrics_computation
virtualenv -p python3 venv
source venv/bin/activate
TMPDIR=$PWD/venv pip install --cache-dir=$PWD/venv -r requirements.txt (--no-deps)

Then the final step is to do the following in a terminal.

cd /path/to/validation_metrics_computation
cp blank_main_config.ini main_config.ini

You can now edit your main_config.ini file for running the different processes.
An additional explanation of all parameters specified in the configuration file can be found in /Utils/resources.py.

3. Process

To run, you need to supply the configuration file as parameter.

python main.py -c main_config.ini

After filling in the configuration file, you should run first the validation task and then the study task.
N.B. If no study fits your need, you can create a new study file in /Studies/.

How to cite

Please, consider citing our paper, if you find the work useful:

@misc{https://doi.org/10.48550/arxiv.2204.14199,
title = {Preoperative brain tumor imaging: models and software for segmentation and standardized reporting},
author = {Bouget, D. and Pedersen, A. and Jakola, A. S. and Kavouridis, V. and Emblem, K. E. and Eijgelaar, R. S. and Kommers, I. and Ardon, H. and Barkhof, F. and Bello, L. and Berger, M. S. and Nibali, M. C. and Furtner, J. and Hervey-Jumper, S. and Idema, A. J. S. and Kiesel, B. and Kloet, A. and Mandonnet, E. and Müller, D. M. J. and Robe, P. A. and Rossi, M. and Sciortino, T. and Brink, W. Van den and Wagemakers, M. and Widhalm, G. and Witte, M. G. and Zwinderman, A. H. and Hamer, P. C. De Witt and Solheim, O. and Reinertsen, I.},
doi = {10.48550/ARXIV.2204.14199},
url = {https://arxiv.org/abs/2204.14199},
keywords = {Image and Video Processing (eess.IV), Computer Vision and Pattern Recognition (cs.CV), Machine Learning (cs.LG), FOS: Electrical engineering, electronic engineering, information engineering, FOS: Electrical engineering, electronic engineering, information engineering, FOS: Computer and information sciences, FOS: Computer and information sciences, I.4.6; J.3},
publisher = {arXiv},
year = {2022},
copyright = {Creative Commons Attribution 4.0 International}}

ingridtv / metrics_computation