Hi,
The website refers to working with binary masks (where the original values are all either 0 or 1). It sounds like you are working with labelled images where the values are no longer binary. In that case there are two options:
- extract separate binary masks for the different areas (e.g., with fslmaths) and then transform, threshold and recombine
- use the interpolation option nearestneighbour, which will preserve the label values and avoid overlaps
The second option is the easiest one to work with, though if the precise borders of the regions are important then the extra control that you get with the thresholds in the first option might be useful (e.g., if you want to make the masks more or less inclusive).
I hope this helps.
All the best,
Mark
> On 14 Jul 2017, at 18:32, Rob Mok <[log in to unmask]> wrote:
>
> Hi there,
>
> I'm using FLIRT to convert anatomical masks (V1 - V3 from the macaque F99 Caret atlas) to individual EPI sessions. One issue is that although in standard space V1 and V2 are separated, once I put them into EPI space they overlap a lot. With experimentation I found that threshold of 4 was good (and actually was 7 even better), so that there was very little overlap between the V1 and V2 masks.
>
> This is confusing to me, since the fsl page (https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FLIRT/FAQ) says:
> a threshold near 1 (say 0.9) is conservative
> • only voxels in the new space that overlap by 90% with the original mask will be included in the new binary mask
>
> This to me means that 1 is 100%. And 4 means 400%.. But should 1 and 4 mean the same thing? I'm guessing the fsl page is just a guideline and it doesn't literally mean this.
>
> Does a threshold of 4 actually make sense? If so I would be happy to use it, but I don't want to use it if it's 'wrong' in some sense.
>
> Thanks,
>
> Rob
>
> code:
>
> flirt -in ${infile} -ref ${regdir}/example_func.nii.gz -out ${outfile} -applyxfm -init ${regdir}/standard2example_func.mat -datatype float
>
> fslmaths ${outfile} -thr ${thrval} -bin ${outfile}
|