Hi,
I am new to using FSL. My aim is to use the sienax script to obtain whole-brain segmentations which I will then convert to a form that is compatible with the software that we use. I have tried this with another brain region using first flirt for registration and FIRST for segmentation and that worked fine.
I have run sienax on a sample batch of scans using the command sienax filename.nii -r -B "-f 0.65 -R -c 104 128 128" with the hope of obtaining whole brain, GM and ventricular segmentations and volumes. I get a whole brain volume estimate value and the following output files:
I2std.mat I.nii I_render.png report.html
I2std.png I_periph_render.nii.gz I_vent_render.nii.gz report.sienax
I_bet.png I_periph_render.png I_vent_render.png
I_masks.png I_render.nii.gz masknz.nii.gz
I am not sure which file is the whole brain segmentation. I thought it was the I_render.nii.gz file, but it would appear that it is an overlay of the segmentation not an actual segmentation itself because when I convert it to a form that is compatible with our software no segmentation comes through.
I then tried it using the -d option of keeping intermediate files and I got the following output:
I2std.avscale I_stdmaskbrain_pve_0.nii.gz
I2std_inv.mat I_stdmaskbrain_pve_0_segvent.nii.gz
I2std.mat I_stdmaskbrain_pve_1.nii.gz
I2std.png I_stdmaskbrain_pve_1_segperiph.nii.gz
I_bet.png I_stdmaskbrain_pve_2.nii.gz
I_brain_mask.nii.gz I_stdmaskbrain_seg.nii.gz
I_brain.nii.gz I_stdmask.nii.gz
I_brain_skull.nii.gz I_stdmask_segperiph.nii.gz
I_masks.png I_stdmask_segvent.nii.gz
I.nii I_vent_render.nii.gz
I_periph_render.nii.gz I_vent_render.png
I_periph_render.png masknz.nii.gz
I_render.nii.gz report.html
I_render.png report.sienax
I_stdmaskbrain.nii.gz
It appears that the I_stdmaskbrain.nii.gz file most closely corresponds to the I_render.nii.gz overlay, but I am not certain if that is a mask or a segmentation.
I would really appreciate some help working out which files are the actual segmentations so I can work out if the problem is the conversion or the segmentation.
Thanks,
Clare
|