search

DevelopmentalImagingMCRI / MCRIBS

Surface version of the M-CRIB atlases.
12 stars 6 forks source link

Surface reconstruction with custom segmentation? #13

Open fleurgaudfernau opened 4 months ago

fleurgaudfernau commented 4 months ago

Hello, I am currently using this pipeline to extract cortical surfaces from fetal MRI data. I would like to segment the fetal images myself and THEN use MCRIBS to extract cortical surfaces. Can I put my own segmentation images in the folder TissueSegMCRIBS and then proceed with MCRIBReconAll --surfrecon? If so, what kind of segmentation outputs are expected by MCRIBS? Thanks in advance, Fleur

chrisadamsonmcri commented 4 months ago

Fleur,

If you just want to use your own labelling, the easiest thing to do is replace the label fusion output file labelfusion_dkt_edited.nii.gz. Ideally, you would use the same labelling scheme as the existing output. There are some postprocessing steps that use this file and others from the pipeline to create the primary file needed by the surface extraction tool regions.nii.gz. This contains labels for left/right WM closed at the midline, cortical GM and brainstem+cerebellum. This file is constructed from the label fusion output and other segmentation images. You could generate this file yourself if you wanted to. You can look at examples of this file from your existing runs if you want to see what you need.

However, there are other files that are generated from the MCRIBS output files that are used in the surface extraction process. These are the wm_force* and pial_force files.

I might need more information from you about what you want to do and I might need to modify the options in the surface extractor to do what you need. Will be happy to do that.

Chris.