Navigation and service

Signal and Image Processing

To reconstruct the nerve fiber architecture from the images generated in the 3D-PLI measurement, various methods of signal and image processing are employed. The methods are validated by analytical methods and numerical simulations.

HPC infrastructure for processing large image data

To process and handle the measurement data of up to 1 Terabyte per brain section, the resources of the Jülich Supercomputing Centre (JSC) are exploited. The developed software utilizes either a large number of processors or graphic cards.

Computational workflowComputational Workflow: This workflow describes the different steps, computing platform and output images of 3D-PLI.
Copyright: Forschungszentrum Jülich GmbH

Methods of Signal and Image Processing

The parameter maps shown in the figure above (transmittance, direction, retardation, inclination) are extracted from the measured raw images by employing different methods of signal and image processing.

Methods of signal and image processingCopyright: Top left/Middle right: adapted from Schmitz et al., Frontiers (2018); Rest: Schmitz/Schober, Forschungszentrum Jülich GmbH