Hi - these platform differences are indeed due to numerical precision
(etc) differences between different platforms and are perfectly
acceptable - particularly as they are so much smaller than the
validated test-retest accuracy of SIENA as a whole (ie median 0.2%
PBVC error on good data). Wrt the final pair, don't forget that you
can't quantitate the difference between two numbers in terms of a %
when the numbers themselves are very small!
Cheers, Steve.
On 3 Aug 2006, at 18:58, Michael Sdika wrote:
> Hi,
>
> I used siena on different machines with different operating
> system : Solaris
> or Linux.
>
> I noticed that I got different results depending of the OS of the
> machine.
> Here is the finalPBVC value for different images for linux and
> solaris :
> -.7792000000 -.7759000000
> -.2740000000 -.2796000000
> -.1023500000 -.0975500000
> -.8798500000 -.8844500000
> -.7798000000 -.7782000000
> -.0671500000 -.0951500000
>
> As you can see the difference can be huge (30% for the last test).
>
> I saw that a similar problem have been reported for FLIRT
> (post 007147 on the 06/03/22) and the siena prblem might be a
> consequence of
> the FLIRT inconsistency.
>
> does anyone have a workaround for this problem or maybe known if it
> have
> been fixed since March ?
>
> Thanks,
>
> Michael
>
>
> ps : I join the siean output file on solaris and linux
>
> <ms0013_01_t1v_to_ms0013_02_t1v_linux.siena>
> <ms0013_01_t1v_to_ms0013_02_t1v_solaris.siena>
------------------------------------------------------------------------
---
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
------------------------------------------------------------------------
---
|