Image reconstruction in a data space for MPI

Abstract

System matrix based image reconstruction in MPI requires the acquisition of calibration measurements, which is a time consuming process during which the imaging device is not available for other measurements. In a clinical setting, the time would be reduced significantly in which the imaging device is available for patient care. In order to enable fast reconstruction algorithms, the system matrix must be kept in main memory, which places special demands on the hardware of the reconstruction computer due to the size of the system matrix. Previously, it has been shown that system matrices may be acquired in dedicated devices which enables off-site calibration measurements. Based on this, we propose a data space for image reconstruction in MPI. In the data space, a network of imaging facilities, calibration and reconstruction experts will be created to facilitate system matrix based image reconstruction. Imaging facilities send measurement data across the data space to the calibration experts, who send back reconstructed images. By integrating reconstruction experts in the data space, new reconstruction techniques can be provided quickly. Following this approach, the scanning devices do not need to take calibration measurements and are always available for patient care. Further, the imaging facilities do not require dedicated hardware for image reconstruction.