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
I'm not sure about the fully continuous expression that they propose there.
With #2657, there will be additional information utilised for streamline-fixel mapping when streamlines have come from FOD segmentation, being the dixel mask per fixel. So as long as the tangent of the streamline-voxel intersection is a part of the lobe that formed that fixel, it will be assigned. But there still needs to be a decision made where that intersection doesn't intersect with any preserved fixel. In that scenario, a weighted attribution of that streamline segment to multiple fixels makes more sense.
The text was updated successfully, but these errors were encountered:
Relevant manuscript:
https://www.frontiersin.org/journals/neuroscience/articles/10.3389/fnins.2023.1199568/full
I'm not sure about the fully continuous expression that they propose there.
With #2657, there will be additional information utilised for streamline-fixel mapping when streamlines have come from FOD segmentation, being the dixel mask per fixel. So as long as the tangent of the streamline-voxel intersection is a part of the lobe that formed that fixel, it will be assigned. But there still needs to be a decision made where that intersection doesn't intersect with any preserved fixel. In that scenario, a weighted attribution of that streamline segment to multiple fixels makes more sense.
The text was updated successfully, but these errors were encountered: