search

ccipd / MRQy

RadQy is a quality assurance and checking tool for quantitative assessment of magnetic resonance imaging (MRI) and computed tomography (CT) data.
https://ccipd.github.io/MRQy/
BSD 3-Clause Clear License
94 stars 28 forks source link

Getting too many nan values for almost the metrics #43

Open gusSCIMOV opened 1 week ago

gusSCIMOV commented 1 week ago

Hi. I'm reaching out to your team from Dr Tiwari's Lab (U Wisconsin-Madison) Thank you for this amazing tool. I'm getting too many nan values for almost the metrics (except for SNR2, SNR3, SNR4, CNR and NUM) for all my files when running MRQy on my brain MRI cohort (nifty files). I was investigating it, and it seems to be related with the overall (mean) metric computation that involves slices with no foreground voxels.

Have you heard about this issue before and how to handle my input files to prevent the pipeline from running on that issue?

image

satishev commented 1 week ago

Can you try running our newer version of the tool here and report back?

Note that the syntax is slightly different since the new tool (RadQy) is intended for both MRI and CT.

gusSCIMOV commented 5 days ago

Can you try running our newer version of the tool here and report back?

Note that the syntax is slightly different since the new tool (RadQy) is intended for both MRI and CT.

Hi, I tried RadQy with the -t option enabled (MRI) but I'm still experiencing the same issue. (my image format is nii.gz) image