-
Notifications
You must be signed in to change notification settings - Fork 0
Description
Hi Korbinian,
I'm using your implementation of Romeo and MCPC3DS.
These are two great algorithms! Thank you so much for providing/maintaining these!
I have the following (possibly odd) use-case:
I have multi-echo gradient-echo acquisitions with bipolar readout.
I use MCPC3DS and ROMEO to estimate coil-specific phase offsets, as well as a field map.
My goal is to then use these in a type of model-based reconstruction, ie. later I try to iteratively fit a complex-valued model to the whole data.
In order to be able to validate the fit, I decided to "leave out" some echo times during fitting.
I can then use the fitted model to predict and compute an error on unseen samples.
However now I am in the situation, where my data has a rather weird (& unrealistic) readout pattern.
Let's say originally we had alternating even (E) and odd (O) echos: "O-E-O-E-O-E".
Now if I keep the third echo time for validation, the new pattern is: "O-E-E-O-E".
I believe this leads to an issue with the bipolar correction & gives worse results.
Here is my question:
Do you think it is possible, to add an argument, such that one can provide the specific readout pattern?
Would this make sense even? Or am I missing something?
Thanks again & best wishes,
Georg