You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When determining image volumes with matched diffusion encoding but opposing phase encoding direction, currently there's just an absolute threshold difference applied to b-values. But if the b-value stored in DICOM reflects the influence of non-DWI gradients, the difference between genuinely matching volumes may exceed this threshold.
The more robust solution is to utilise the same b-value clustering algorithm as is utilised elsewhere to group image volumes into shells, and then ensure during the volume matching process that the two volumes are assigned to the same shell.
The text was updated successfully, but these errors were encountered:
- In dwifslpreproc, when determining whether or not explicit volume recombination should be performed (as well as which pairs of volumes should be combined), rather than using a hard-coded b-value difference threshold, instead obtain the outcomes of applying the C++ shell clustering algorithm to those data, and ensure that paired volumes are attributed to the same shell.
- mrinfo has a new option -shell_indices to facilitate this.
Closes#1848.
Following from forum thread.
When determining image volumes with matched diffusion encoding but opposing phase encoding direction, currently there's just an absolute threshold difference applied to b-values. But if the b-value stored in DICOM reflects the influence of non-DWI gradients, the difference between genuinely matching volumes may exceed this threshold.
The more robust solution is to utilise the same b-value clustering algorithm as is utilised elsewhere to group image volumes into shells, and then ensure during the volume matching process that the two volumes are assigned to the same shell.
The text was updated successfully, but these errors were encountered: