Albo is a tool for calculating the overlap between a lesion and brain regions, based on MRI sequences. Given a set of MRI images, the lesion is automatically segmented and then registered to a standard brain. Now, the lesion overlap with brain regions is calculated based on brain atlases.
The following programs are needed for installing and running albo:
- NumPy
- SciPy
- Computational Morphometry Toolkit (CMTK)
- FSL
- NiftyReg
Apart from NiftyReg, in Ubuntu these programs can be installed with the command
sudo apt-get install python-numpy python-scipy cmtk fsl
NiftyReg needs to be built from source, see here for instructions.
After installation, you may need to add something like source /etc/fsl/5.0/fsl.sh
to your .bashrc file, depending on your FSL version. This
makes, e.g., the command bet
available, instead of fsl5.0-bet
, which is
required by albo. See man fsl
for details.
Download and unpack the program (or clone the repository) and run python setup.py install
. This may download further Python dependencies.
Open the file ./config/albo/albo.conf
and set the paths needed by the program.
Then, place the classifiers, standardbrains and
This section provides just a rough overview. See the
wiki for more
information, or run albo -h
in your terminal.
Configure the necessary paths in ~/.config/albo/albo.conf
.
To list all available classifiers and the sequence identifiers they emply, use
albo list
Assuming there is a classifier identifying the FLAIR sequence as MR_Flair and the T1 sequence as MR_T1, to run the program, use
albo run --id case-id MR_Flair:path/to/flair MR_T1:path/to/t1
In case new atlases were added to the atlas directory, re-calculate the overlap for all previously segmented cases with
albo update