Hi,
My guess is that the image is of an integer type. I believe fslmaths will
process everything as a floating point number then output to the original
data type unless otherwise specified. If you append "-odt float" to your
command line, it should resolve your issue. -odt specifies the output data
type.
cheers,
Brian
P.S. You can use `fslinfo image.nii.gz` to check the image data_type
> Hello,
> I'm having an issue with fslmaths and am just curious if this is expected.
> I have a set of mouse brains with values in the range of 8000-11,000. As
> we used a contrast agent we want to be able to normalize the values by
> dividing the whole thing by the intensity of the pituitary, which has a
> general range of 12,000-20,000.
>
> If I use fslmaths with this command line (01_flirt is the brain. 19428.26
> is the pituitary intensity) I end up with a binarized image with values of
> 1 throughout.
> fslmaths 01_flirt -div 19428.26 01_flirt_pit
>
> I thought possibly it was the decimal, but this yields the same thing, as
> does replacing 19428 with 10,000.
> fslmaths 01_flirt -div 19428 01_flirt_pit
>
> To test it, I made the number to be divided by smaller by a factor of 10
> (so, 1942.8), resulting in an image with a range of about 1-11. If I
> reduce it by a factor of 100 (194.28), the output has a range of about
> 1-60.
>
> Is there something wrong with my approach or am I missing something? If
> not, is there some way to get it to just do the math without
> truncating/rounding/binning the output?
>
> any advice is appreciated!
> thanks!
> Katie
>
>
> _________________________________________
> Katherine H. Karlsgodt, Ph.D.
> Center for Cognitive Neuroscience
> Department of Psychiatry
> University of California, Los Angeles
>
> email: [log in to unmask]
> phone: 310-206-3019
> fax: 310-794-9740
>
>
>
>
>
>
>
|