2025/03/14 by Nawwas, Lina, Möddel, Martin, Knopp, Tobias
#Matrix Compression #Multi-contrast MPI
paper · doi:10.18416/ijmpi.2025.2503062
Multi-contrast magnetic particle imaging (MPI) reconstructs the signal from different tracer materials or environments, resulting in multi-channel images that enable temperature or viscosity quantification. Since the multi-contrast problem is ill-posed, it is addressed by regularization methods that are commonly solved using the Kaczmarz algorithm. Unlike the single-contrast MPI problem, the multi-contrast one requires a high number of iterations to converge. Matrix compression techniques were already successfully used in single-contrast reconstruction and matrix recovery applications as in compressed sensing. Our work proposes to use matrix compression to reduce the reconstruction time needed to achieve good reconstruction quality in multi-contrast MPI.