Hi Jack,
>This still isn't quite clear to me. Our physicist wrote a program to do the following:
>1. Acquire 2 gradient echos with a 10ms interval and
>2. Determine how much phase has been acquired during those 10ms
>
>As I understand it, this gives you the "phase map" and its intensity is proportional to
>the inhomogeneity in the field. In fact, the output of his program looks a lot like the
>phase maps that you guys have shown at the FSL course.
>
Yes, this difference will look like the field map. I'm afraid the
problem is what we are
each defining as a "phase map". I haven't been very strict in my
terminology in fugue
and so the input asks for a "phase map" but I actually mean a 4D image
containing a
pair of unwrapped phase images (one per TE or asym setting) which will
be turned into
a phase map and then a field map internally. Sorry for the confusion.
>The input to Prelude is two images: one giving the phase and one giving absolute
>values. The output is a real, unwrapped phase image.
>
Yes, prelude is very simple. Input = complex image, output = phase
image (real valued).
Alternative options for specifying the input include (1) a fully complex
image (avw datatype 32)
or (2) a pair of real images, one representing phase, the other
representing absolute values
or magnitudes.
>So, why would there be two images as the input of Fugue when there is one image
>as the output of Prelude? When you talk about the "two phase images", are you
>referring to the two gradient echos? Is Fugue looking for real, imaginary, or complex
>input as the <unwrapped phase map>?
>
>
OK, prelude will unwrap as many input images as appear in the input
file. So if the input file
contains 2 complex images (as a 4D complex analyze type) then the output
will also be a 4D
file containing 2 images (respective unwrapped versions of the input
ones). This is the
default behaviour for prelude: input = one 4D file containing 2 lots of
3D images. The
2 images are at different asymmetries (first is fully symmetric, second
has asym > 0).
For gradient echo sequences they would just be different echo times, yes.
For the <unwrapped phase map> fugue requires one 4D analyze file
containing 2 lots of
3D real, unwrapped phase images.
I hope that answers all of your questions and is clear enough.
Once again, sorry for the confusion.
All the best,
Mark
|