 |
 |
Hello, I had a look at how we could carry out a SENSE reconstruction with gadgetron. I found two examples one for iterative (conjugate gradient or others) SENSE (cpureg_cartesian_averaging.xml) and one for GRAPPA (Generic_Cartesian_Grappa.xml). They use different types of pipelines for the reconstruction. Here is how I understand it: cpureg_cartesian_averaging.xml 1) CartesianToGenericGadget creates an additional hoNDArray which contains the trajectory and the dcf (simply 1s for the Cartesian case). This is required because the following gadgets are implemented for non-Cartesian trajectories. 2) gpuGenericSensePrepGadget adds k-space data used for the csm estimation to a buffer and then reconstructs images and obtains the csm information prior to the iterative SENSE reconstruction and creates a GenericReconJob object which contains all the required information for the reconstruction (k-space data, trajectory, dcf, csm, regularisation,...) 3) gpuCgSenseGadget carries out the interative SENSE reconstruction. Other iterative SENSE reconstructions are also possible. Generic_Cartesian_Grappa.xml 1) GenericCartesianReconReferencePrepGadget receives an IsmrmrdReconData object which contains the k-space data for the image and also for the csm estimation (no need for trajectory or dcf information because it only works for Cartesian) 2) GenericCartesianGrappaReconGadget carries out the GRAPPA reconstruction For the first case I guess we would require an interface to the GenericReconJob object to add the estimated csm information before carrying out the gpuCgSenseGadget. In the second case we would need to modify the IsmrmrdReconData object to include our external csm information and modify the GenericCartesianGrappaReconGadget to do a SENSE rather than GRAPPA reconstruction. Which of the two ways do you think we should try first? Or would it make more sense to write a simpler gadget ourselves rather than using the more general (and therefore complex) gadgets which already exist? Regards, Christoph -- Dr. Christoph Kolbitsch Medical Physics and Metrological Information Technologies Physikalisch-Technische Bundesanstalt Abbestr. 2-12 10587 Berlin, Germany phone: +49 30 3481 7761