http://imagej.273.s1.nabble.com/StackReg-or-similar-tp5000226p5000246.html
> It Contains:
>> - s03_SOURCE.tif
>> - s03_StackReg.tif (very fast and good)
>>
>> (not good/fast enough (maybe there are to me unknown tweaks))
>> - s03_ImageStabilizer.tif
>> - s03_RegVirtualStacks_Rigid.tif
>> - s03_RegVirtualStacks_Transformation.tif
>>
>>
>> I think there are purposes, when the actual registration with its
>> sampling is not needed and could be disabled to only get transformation
>> data for speed reasons.
>>
>> Regards,
>> Rainer
>>
>
> I liked playing a bit with the data set with a macro I wrote: my thinking
> was to make a macro which first makes a stack average (z-project), then
> aligns the whole stack to that, iterates until some convergence level is
> met. I think, though, that this process can get stuck in a local minimum,
> say if there are two roughly equivalent registrations, and I am not sure
> how to avoid that. One thing I think that is missing from stackreg is the
> ability to align all images to one reference image, so that the stack does
> not drift over time. So actually, I just tried stackreg first, then my
> iterative macro, and it works great! I won't attach it, as it is ~3MB, but
> it has both no drift and excellent image-to-image registration. I'll send
> it individually if you want.
>
> JPK
>
>
>
>
> *******************************************
> Jacob Pearson Keller
> Northwestern University
> Medical Scientist Training Program
> email:
[hidden email]
> *******************************************
>