Hi,
The robust range is more complicated than this, as it tries to capture the a representative range even when the distribution is tricky, such as when it is a binary mask with only a few non-zero voxels. There is a rather involved algorithm used to calculate this robust range, though for typically distributed data it is normally similar to the 2nd and 98th percentiles.
The other two are correct.
If you just want the 2nd and 98th percentiles that include the zeros then "-p 2 -p 98" will do the trick.
Also note that this syntax is for fslstats, not fslmaths as you have written.
All the best,
Mark
On 13 Jan 2014, at 15:40, Varina Wolf <[log in to unmask]> wrote:
> Hello Experts,
>
> Please let me know if I have this definition correct
>
> fslmaths image -r = robust range = the 2nd to 98 percentile, including zeros
> fslmaths image -R =absolute max and min
> fslmaths image -P 98 = 98% percentile, not including zeros
>
>
> Thank you!
> Varina
|