Quantitative MRI and feature extraction from MR images typically requires large continuous optimization problems to be solved. Alternatively, many smaller problems are sometimes solved, allowing for easy parallelization. The aim of the project is to estimate physical quantities from a set of MR images. The many voxels involved makes the resulting problems large, both in terms of number of parameters, and the number of data points.
The obtained estimates increase the information basis, and can be used by medical doctors to improve diagnosis.