Unexpected FANSI results with ex vivo data #114
Description
Hello,
I'm trying to perform QSM using FANSI and have encountered some unexpected results.
I'm working with an ex vivo spinal cord MRI image with dimensions (64×64×1×8), intended for histology comparison. After performing phase unwrapping with ROMEO, I replicated the same image 5 times to create a "3D image" compatible with SEPIA (SEPIA does not seem to work with 2D images). I then compared three background field removal methods: VSHARP, PDF, and LBV (RESHARP and SHARP do not seem to work with my data).
I've applied several QSM methods from SEPIA, including Closed Form Solution, iLSQR, TKD, and FANSI. While the susceptibility maps obtained from the other methods are not identical, they are quite similar. However, the output from FANSI is significantly different from all the others.
Do you have any idea why FANSI might produce such results?
Thank you very much for your time.
Alexandre B