MRE-J: A novel image processing pipeline for Magnetic Resonance Elastography using ImageJ and Apache Commons-Math

Magnetic Resonance Elastography directly measures tissue stiffness of an image volume in vivo. The four stages of MRE image processing – unwrap, filter, decompose, and invert – are accomplished by widely varying procedures depending on research group. Additionally, related image processing procedures in neighboring fields have not had their potential investigated. A new image processing pipeline was created using ImageJ with matrix support from Apache commons-math. Various choices for each step of the pipeline were implemented to allow for comparison between methods of various groups, as well as to implement potentially applicable techniques from other disciplines. Our own pipeline was selected after testing the available choices. The results allow for high anatomical specificity in the elastogram without interpolation, null filling, or other manipulations and without leaving the resolution of the original pixel matrix. Future developments include the addition of viscoelasticity calculation.