[feature] add subspace projection model. #201
Conversation
| # installable using ``pip install brainweb-dl``. | ||
|
|
||
| from mrinufft.extras import get_brainweb_map | ||
|
|
There was a problem hiding this comment.
I am not sure I got this - do you think we should add a comment specifying we are selecting the 90-th slice?
There was a problem hiding this comment.
just slice selection for 2D processing
There was a problem hiding this comment.
Hello Matteo, Thanks for this PR, having the test examples and documentation altogether really helps to grasp the new feature.
I have some minors comments + I would advise to have the dimension ordered as B,T,C,... as the batch dimension is for independent data points. (but if it makes more sense for the computation to be done in T,B,C... you can just move the axes around)
|
I am not sure why tests would fail for gpunufft density compensation - I did not touch anything from that subpackage |
|
The CI is just a bit flaky it seems. |
This PR add a new extension to enable low rank subspace projected NUFFT as described (in a Cartesian context) here. A non-exhaustive list potential applications is: