[Teem-users] Re: txf bake-off

[Jon M. <jon...@mo...>]

On Wed, 2005-03-30 at 21:10, Gordon Kindlmann wrote:
> Fixing bane has been on my todo list 
> for so long its sad.  It wouldn't be that hard.  But, if you're a real 
> masochist, you can get the guidance from value/gradient scatterplots 
> created via vprobe and then "unu jhisto", and then based on these you 
> can use "gkms txf" to make transfer functions.  I can give you some 
> more pointers on this if you're interested (ask again on teem-users).
> I have some hack to get gkms to compile even though bane is messed- the 
> "txf" command is actually independent of bane so it works.  This is 
> then something you can feed into miter.

Yes I'd like to try it out, so let me know how your work-around works.

> One thing to keep in mind is that if you have a lot of 
> datasets of the same modality, collected with the same parameters, of 
> the same kind of specimen (common in science), one txf may work for 
> them all, which lessens the needs to do clever semi-automatic transfer 
> functions.

This is largely true, but with biological (in vivo) systems the
variability across subjects can be surprisingly high even when modality
and acquisition parameters are identical.  

--
Jon