Hi,
This could be an issue with the first build of 415 - an updated build
was released today which should fix this..
Many Regards
Matthew
>
> I just noticed a problem with FEAT output images - post stats images
> appear to have no activation, due to a LUT error. All above-threshold
> voxels appear white instead of having the appropriate heat map applied.
> I believe this is due to a change/bug in slicer:
>
> slicer rendered_thresh_zstat1.nii.gz -d -S 2 750
> rendered_thresh_zstat1.png
> Option 0: -d
> Option 1: -S 2 750 rendered_thresh_zstat1.png
> -S 2 750 rendered_thresh_zstat1.png
> Entries in LUT: 0
> Intensity min, max : -4.03287 3.23551
>
>
>
> Doesn't find any LUT. Some of my rendered* files appear to have
> render1t specified in their header's aux_file while others do not
> (version dependent??) but this does not seem to help slicer find the
> correct LUT. Manually specifying -l render1t correctly loads the LUT
> but does not scale it properly - intensity min, max are set to the
> [0.02, .98] percentile range instead of the header cal_min, cal_max.
>
> Changing featlib.tcl as follows:
>
> fsl:exec "$FSLDIR/bin/overlay $fmri(rendertype) 0 $underlying -a
> thresh_$rawstats $fmri(zmin) $fmri(zmax) rendered_thresh_$rawstats"
> # Had to modify this - SG
> #fsl:exec "${FSLDIR}/bin/slicer rendered_thresh_$rawstats $SLICER 750
> rendered_thresh_${rawstats}.png"
> set iminmax [ fsl:exec "${FSLDIR}/bin/fslstats rendered_thresh_$rawstats
> -R 2>/dev/null" ]
> set imin [ lindex $iminmax 0 ]
> set imax [ lindex $iminmax 1 ]
> fsl:exec "${FSLDIR}/bin/slicer rendered_thresh_$rawstats -l render1t -i
> $imin $imax $SLICER 750 rendered_thresh_${rawstats}.png"
>
>
>
> Appears to fix the problem. I am running 4.1.5 from the
> neuro.debian.net repository on Ubuntu 9.10, tried in both 32 and 64 bit
> machines (i have also tried directly compiling those sources, and that
> does not help). I am not positive when the above broke as I upgrade
> frequently and do not always look at FEAT output (as I usually do
> further analysis or view on the surface via Freesurfer), but slicer has
> rendered correctly on my machines in the past
>
> Thanks,
> -Scott Gorlin
>
>
|