3D Registration of Stacked Sections #55
Description
Hi,
Thank you for put out this very nice package. I was hoping you could help answer a couple challenges I've been having when trying to run this for registration of our MERFISH dataset.
I have a 3D stack of MERFISH sections from the same mouse brain that I'm trying to register together into the Allen CCF (v3). And I am running into the following problems:
-
The tutorials seem to be focused around registering a single section to an approximate slice in the CCF, but it is clear from our data that there is a global obliqueness mismatch between our mouse brain and the CCF atlas brain (specifically, a skew in the A/P direction). Is there some way to perform an initial 3D affine alignment of the sections relative to the atlas, so that the MERFISH sections can be matched with oblique cut-throughs of the CCF? If I just insert a single section as suggested in the tutorial, it doesn't seem to be able to capture this skew relative to the CCF and cause quite a bit of structural mismatch.
-
I was also wondering if there is maybe some way to enforce a certain level of coherency between results across different sections. If I register each section one at a time into the CCF, unsurprisingly, the registered sections lose their structural coherence with each other, leaving large jumps/discontinuities between the same structure boundaries across sections. It would be helpful if there was a way to enforce some kind of smoothness between sections or some kind of a penalty on changes to the 3D continuity.
-
Similarly, is there a way to limit the individual section results so that they don't intersect each other in the CCF after registration?
-
Lastly, a more trivial issue. I can't seem to figure out how to use the transformation result and apply it to overlapping image data in the same space as the MERFISH data. Specifically, I'd like to apply the transforms to anatomical labels maps of the MERFISH data and move it into the CCF.
Any help would be appreciated!
Thank you,
~Min