I wouldn't suggest changing the headers. The values, which appear to
be floating point, are actually from uint8 values in the range of 0 to
255, which have been scaled by the scalefactor of 0.0039 in the NIfTI
header.
Best regards,
-John
On 29 October 2012 23:20, Cuneo, Daniel P. <[log in to unmask]> wrote:
> Hi all,
>
>
>
> My lab uses SPM8 segmentation to generate Gm, Wm and Csf seg images (
> intensity range 0.0 – 1.0 ).
>
> We expect to see the output data type as float, but it is uint8…? We have
> been using spm8 segmentation for at least a year,
>
> and never noticed this before. Most viewers still show the data accurately
> and MATLAB of course loads the data as double.
>
>
>
> However we can’t leave images with inaccurate headers laying around…I know
> how to edit the headers, but wanted to
>
> check if this is a known bug or something left out of the configuration ?
>
>
>
> TIA,
>
> dpc
|