The loopUI Python package provides several functions to analyse the variability among an ensemble of voxets. It allows the computation of local measures of uncertainty (cardinality and entropy) and of global dissimilarity measures (e.g. based on multiple-point statistical analysis or wavelet decomposition among other possibilities).
Note: package for python >=3.9
To install the package: python -m pip install loopui
To uninstall the package: python -m pip uninstall -y loopui
The following python packages are used by 'loopui':
- matplotlib
- mpl_toolkits
- numpy == 1.26.4
- gzip
- scipy
- sklearn
- pywt
- pandas
A series of Python notebooks demonstrate how to use the different functions.
First, run ui-0-data.ipynb
to load an ensemble of voxets. It is a pre-requisites prior running the other python notebooks.
In order to compare the various indicator via the viewing ui-9-comparison.ipynb
notebook, a prerequisite is to run the cardinality and entropy notebooks (other pre-computed indicators are saved in the pickledata folder, but the indicators can be computed by running the appropriate notebook).