Liom Toolkit
This package supports the research being done by the Laboratoire d’Imagerie Optique et Moléculaire at
Polytechnique Montréal. It hosts a collection of scripts used to process and analyze data collected by the lab.
Installation
The package can be installed using pip:
pip install liom-toolkit
Due to the complicated requirements, a detailed installation guide is provided below.
Usage
Demonstrations of some of the functionalities of the package can be found in the notebooks in the LIOM Notebooks repository.
The repository can be found here: LIOM Notebooks
Requirements
The package requires the following packages to be installed and will attempt to install them using installation:
- antspyx
- tqdm
- scikit-image
- ome-zarr
- nibabel
- zarr
- h5py
- pynrrd
- PyWavelets
- SimpleITK
- allensdk
- dask
- opencv-python
- torch
- torchvision
- wandb
- patchify
- natsort
- albumentations
To create an anaconda environment with all the required packages, run the following commands:
conda create -n <name>
conda activate <name>
conda install python=3.10
conda install pytorch torchvision torchaudio pytorch-cuda=12.1 -c pytorch -c nvidia
conda install pytorch::pytorch torchvision torchaudio -c pytorch
brew install hdf5
HDF5_DIR=/opt/homebrew/Cellar/hdf5/1.14.3_1 pip install tables
pip install allensdk
pip install antspyx
pip install liom-toolkit
pip install sphinx-rtd-theme
pip install sphinxcontrib-apidoc
pip install jupyter
Package Structure
The package contains the following modules:
Registration
The registration module is concerned with performing registration on brain imagery. It hosts a collection of scripts for
registering mouse brains to the Allen Atlas as well as functions for creating brain templates to use in registration.
Segmentation
The segmentation module is concerned with segmenting brain imagery. It contains methods for segmenting brain images into
different regions of interest. The vseg submodule contains methods for segmenting vasculature using deep learning using
a U-net architecture. The pretrained model is trained on LSFM data.
Utils
Various utility functions used by the other modules. These include function for converting between the different data
files used within the lab.