How to use dcmri#

Analyzing data#

To analyse data with dcmri, select the appropriate tissue type from the tissue bank and train it on your data. For examples of usage, consult the documentation of the model or look for a similar application in the list of examples.

The models in the tissue bank are high-level implementations that can be customized to run a wide range of different models, parameter settings or methods. If no configuration options are specified by the user, they will always run the most conventional models. For more background on what models are available, or what their parameters are, have a look at the relevant background section.

Developing models#

Apart from the end-to-end models in the tissue bank, dcmri also includes a library of more generic basic methods that can be used to build custom-models more easily, and facilitate the creation of new models to extend the functionality of dcmri. These basic methods are implemented as simple python functions for maximal transparency and modularity.

The are organised in a hierarchical fashion:

  • The tissue module contains high-level implementations of tissue-specific functions such as concentrations or signals. For more background, see the section on tissue types.

  • The signal module provides generic signal models for MRI sequences. For more background on these models, see the section on imaging sequences.

  • The pharmacokinetic building blocks implement generic pharmacokinetic models that can be assembled to build more complex models. For more background on these models, see the section on basic pharmacokinetics.

  • A library of utilities that can be used to build new functions, test them or demonstrate their usage. This includes:

    • A library of real data taken from published studies. These are all available through the dcmri.fetch function (see also section examples).

    • A library of functions to generate synthetic data for testing or demonstration of models.

    • Synthetic images the can be used to build digital reference objects.

    • A library of published input functions.

    • A library of useful constants taken from literature, such as standard dosages, concentrations and relaxivities of common contrast agents, and common MRI parameters and perfusion parameters for different field strengths or tissue types.

    • A collection of basic functions for performing convolutions, or sampling data.