search

GSTT-CSC / hazen

Quality assurance framework for Magnetic Resonance Imaging
https://github.com/GSTT-CSC/hazen
Apache License 2.0
19 stars 4 forks source link

[HZN-66] SpRes-should account for magnitude images #66

Open laurencejackson opened 2 years ago

laurencejackson commented 2 years ago

Transferred by @LaurenceJackson on Aug 9, 2020, 18:23

:::Issue recreated from Jira issue HZN-66:::

Imported from https://bitbucket.org/gsttmri/hazen/issues/67/spres-should-account-for-magnitude-images The images we acquire are generally magnitude images. The test images in Hazen appear to be magnitude images also. This causes a problem when generating the signal in the void as there is often a phase change which needs to be accounted for. The current code does not do this. It can be observed in the attached image.The ERF between c.62-68 has been rectified (by the scanner’s magnitude recon). It should be negative (and at each further wavelength). This can also be seen in the LSF—the spatial frequency on the right of the line is roughly double that on the left.This could be addressed by requiring a complex image (or at least signed) acquisition, or by rectifying the signal in post processing.There is useful information on the pitfalls of this technique in:Steckner, M.C., Drost, D.J. & Prato, F.S., 1994. Computing the modulation transfer function of a magnetic resonance imager. Medical physics, 21(3), pp.483–489. https://europepmc.org/article/med/8208224

RFranklin-MR commented 10 months ago

Physics review - assumed low priority as results with Magnet phantoms seem to be as expected.