search

imr-framework / mrf

An end to end, vendor neutral MRF package
GNU Affero General Public License v3.0
31 stars 11 forks source link

the slide windows #3

Closed ghost closed 5 years ago

ghost commented 5 years ago

I noticed that you reconstruct the images by using 48 or partial trajectory data,but the Nature paper "MRF" said the reconstruction only need a single trajectory data,so what's the reason of your solutions?

qianenlin commented 5 years ago

Thank you for posting the question!

The MRF sequence designed in Sequence_Design part uses 48 shots of spirals. So by setting up the size of sliding window as 48, the selected trajectory will cover the full k-space, which will improve SNR and robustness of dictionary matching.

A good reference will be Cao et al., 2016, please let me know if you have any questions!

qianenlin commented 5 years ago

@ouyangbinyu May I ask if I answered your question? If so, can I close the issue? Thank you!