search

daviddmc / NeSVoR

NeSVoR is a package for GPU-accelerated slice-to-volume reconstruction.
MIT License
69 stars 16 forks source link

Can't get the full volume using actual thickness. #20

Open ArnoBlue opened 3 months ago

ArnoBlue commented 3 months ago

Hi NeSVoR team,

Thank you so much for making the code public! I tried to reproduce your results using the CRL atlas dataset. But found that I need to enter a larger parameter than the actual thickness to get the full volume.

I used the following code to get to the thickness of the data as [0.8,0.8,0.8]. img = nib.load(nii_file) #turn into numpy data = img.get_fdata() affine = img.affine resolution = affine.diagonal()[:3]

I used the following command for SVR：

nesvor reconstruct \ --input-stacks .../STA33.nii.gz .../STA33.nii.gz .../STA33.nii.gz \ --thicknesses 0.8 0.8 0.8 \ --output-volume .../3316.nii.gz \ --bias-field-correction \ --output-resolution 0.8

However, I got an incomplete volume：

image

If I change the command to --thicknesses 1.4 1.4 1.4 \ I can then get the full volume:

image

ArnoBlue commented 3 months ago

I have noticed that the results I get are similar to those in 'Cannot reproduce results on FeTA using your slicing strategy' (a closed issue). Could you clarify what are the limitations of the thickness parameter?