search

sct-pipeline / fmri-segmentation

Repository for the project on automatic spinal cord segmentation based on fMRI EPI data
MIT License
4 stars 1 forks source link

Experiment: Inference of a 3d nnUNet on a single slice #20

Open rohanbanerjee opened 1 year ago

rohanbanerjee commented 1 year ago

During a discussion with @kennethaweberii and @MerveKaptan, we discussed if our 3d automatic segmentation model would work if it has to segment the spinal cord on a single slice.

I am currently investigating this and will update this issue with the different outcomes that I get.

jcohenadad commented 1 year ago

from ROHAN: Sct_image padding doesn’t work for a single slice. Throws the error “ERROR: you need to specify a 3D input file.” --> TODO: open issue and crossref

jcohenadad commented 1 year ago

from ROHAN: I wrote a script to pad the images but the nnUNet still throws an error of not being able to segment the images. Investigating this now (Will update all of this on the issue).

rohanbanerjee commented 1 year ago

Sct_image padding doesn’t work for a single slice. Throws the error “ERROR: you need to specify a 3D input file.” --> TODO: open issue and crossref

The issue related to this comment can be found here: https://github.com/spinalcordtoolbox/spinalcordtoolbox/issues/4223#issue-1906112487

I wrote a script to pad the images but the nnUNet still throws an error of not being able to segment the images. Investigating this now (Will update all of this on the issue).

The issue related to this comment can be found here: https://github.com/spinalcordtoolbox/spinalcordtoolbox/issues/4222#issue-1905732792