Global presentation
Brain-Mapper is the product of a 5-month collaboration between the Central Hospital of Nancy, France (CHRU in french) and Computer Science and Engineering students at TELECOM Nancy
The aim of this project was to simplify the treatment chain of neurosurgeon Dr. Fabien Rech, who had to use several librairies, software and methods to extract interesting data from NIfTI files.
The workchain of NIfTI processing and information extraction carried out by Dr. Rech for his thesis work was the following one :
Our aim was to simplify this work chain and offer the possibility to manage NIfTI, extract the interesting data from them to use it for calculations and clustering algorithms
Software usecase
This software is written in Python. Its user interface is written in PyQt 4.
It can be used to create several sets and image collections associated with brain responses or patients, that will contain different images, in the NIfTI format and standards.
The user can then extract data from the loaded NIfTI files to perform basic calculations or apply clustering algorithms to this data.
Calculations and clustering results can be exported as NIfTI or CSV files, as well as the original NIfTI files. Data can be imported as well as NIfTI or Excel files and the user can specify
a workspace directory from which files will be automatically imported
WARNING
This software was designed for the processing of binary NIfTI files, i.e. NIfTI files where voxels have an intensity either equal to zero or equal to 1.
Data extraction is still quite naive : our algorithms extract all points that have an intensity value greater than zero. Thus, if you have NIfTI files with the MNI template
still as a background and your selected regions, our software will extract a list containing all coordinates that have an intensity value other than zero.
We are currently working on a version of the software that allows the user to specifiy a threshold for data extraction but for now, it will be best to use only binary NIfTI files.
Otherwise, extraction time and performance may hinder the utility of the software
Thus, the current version of this software allows the user to :
- Open NIfTI files ('.nii' or in compressed format '.nii.gz') and put them in
Sets, which are associated with a brain response and Image Collections, associated mostly to patients
- Import and export the workspace from/into a directory, whose structure will reflect that of the application (set folders, image collections subfolders...)
- Edit the content of sets and image collections, and change their names
- Visualize in a 3D graph the contents of a NIfTI image from the edition page
- Select several image collections, across all sets, containing the images to which calculations and clustering algorithms will be applied
- Perform several calculations on selected NIfTI files
- Extract data from selected NIfTI files and perform clustering methods on this data (KMeans, KMedoids, AgglomerativeClustering)
- Visualize cluster assignment proportions and samples silhouette through graphical representations in clustering page
- Calculate internal validation indexes for clustering results, such as Silhouettes, Mean Silhouette, Calinski-Habaraz score and Davis-Boulin index
- Save clustering and calculation results as sets and image collections within the application
- Export clustering assignments data table as a CSV file and clustering validation indexes as a Text file
- Export whole image collections or sets as NIfTI files
INFORMATION
Our software can open compressed NIfTI formats ('.nii.gz').
Install
brainMapper is written in
Python 2.7. In this section, we present a tutorial to use the latest version of brainMapper in your computer
WARNING
The desktop application of this software is not available yet
This install tutorial will allow the user to easily execute brainMapper but will not install a compiled version of it on the machine.
We will include compiled desktop applications in the near future.
Dependencies
Our software uses the following Python libraries :
- PyQt 4 (>= 4.0.0)
- NiBabel (>= 2.0.0 )
- Numpy (>= 1.8.2)
- scikit-learn
- pyqtgraph
- pyopengl
Ubuntu (GNU/Linux)
- Open a terminal and install all dependencies with the following commands :
sudo apt-get install python-qt4
sudo python2 -m pip install nibabel
sudo python2 -m pip install -U numpy
sudo python2 -m pip install -U scikit-learn
sudo python2 -m pip install pyqtgraph
sudo python2 -m pip install pyopengl
- Clone the repository or download a .zip file from this project's GitHub repository
and unzip it in a folder of your choice.
- Create a .sh file next to the extracted folder that contains the following commands :
cd brainMapper-master/UI
python brainMapper_UI.py
- Create a shorcut to this .sh file and place it where you want to : this file will be the application's "launcher"
OSX (Mac)
Windows
- Install Miniconda package manager for Windows and Python 2.7: Download the installer here.
When installing, tick the option "Add Anaconda to my PATH variable".
- In a Powershell terminal, type the following commands :
conda install python=2.7
activate root
conda install -c anaconda pyqt=4.11.4
conda install numpy
conda install scikit-learn
conda install pyqtgraph
conda install pyopengl
conda install -c conda-forge nibabel
- Download the .zip project on the GitHub repository.
Decompress it, and create a .bat file (be careful with file extensions on Windows, sometimes a .txt will be added and you cannot see it.).
- In the .bat file put in the following commands
cd brainMapper-master/UI
python UI.py
This file has to be in the same folder than the decompressed .zip.
Finally, create a .bat shortcut (right click, 'Create shorcut') and put it on your Desktop : when you click on it, brainMapper will launch .
