Rician likelihood loss function for quantitative MRI with deep learning

This repository contains a walkthrough and example implementations of the Rician likelihood loss function for self-supervised quantitative MRI (qMRI) as described in Parker et al (2023) [1].

The walkthrough goes through a simple example of applying the Rician loss function for self-supervised qMRI. The example implementations provide syntax to specify and call the Rician loss function in PyTorch, TensorFlow and Keras.

Note that the walkthrough is an adaptation of the demo hosted at https://github.com/sebbarb/deep_ivim - many thanks to the authors for making their demo code easy to use and openly available.

References

[1] Parker CS, Schroder A, Epstein SC, Cole J, Alexander DC, Zhang H. Rician likelihood loss function for quantitative MRI with deep learning, arXiV.