Help with Lobe Segmentation and 3D Visualization in FreeSurfer

💬 Support & Discussions
1

I’m working on segmenting and visualizing the cerebral lobes from T1-weighted MRI scans using FreeSurfer. My objective is to:

  1. Segment the cerebral lobes from T1 MRI.
  2. Assign each lobe a different label and visualize them in 3D.
  3. Register the segmented lobes onto the original T1 scan for clear visualization.
    I followed these steps in FreeSurfer:
    Ran the full FreeSurfer pipeline:
    recon-all -s my_subject -i T1.mgz -all
    Converted the parcellation to individual labels:
    mri_annotation2label --subject my_subject --hemi lh --annotation aparc.DKTatlas --outdir labels/
    2.Merged the labels into lobe-specific segmentations:
    mri_mergelabels labels/lh.    .label -o lh_lobes.label
    Converted the label files into volumetric images:
    mri_label2vol --label lh_lobes.label --temp T1.mgz --o sujet1t1.nii --identity

The Problem:

  • The output does not display as expected. Instead of solid, colored lobes, I see only contours outlining the regions
    Capture d’écran 2025-03-08 134006
    Capture d’écran 2025-03-08 1340061213×877 77.9 KB
  • Applying --apply_lut does not seem to fix the issue.
  • I tried mri_binarize to ensure the labels were filled, but it didn’t change the visualization much.

Does anyone have experience with this type of segmentation? How can I ensure that each lobe is properly labeled, filled, and visualized in 3D correctly? Any help would be greatly appreciated!

Thank you in advance