CNMF-E not segmenting out visually obvious cell ROIs

[camilledonoho]

Hi,

I have tried many many parameter sets on my videos and find that, even when I am incredibly lenient with my parameter choices (choosing parameters that result in 400+ segmented ROIs), some cells that look very obvious by eye are still not segmented/circled. My videos are of somewhat uneven brightness, such that half of the field of view has a brighter baseline than the other half. This likely is part of the problem as the un-circled cells are all in the darker half. Still, the cells at their brightest are still much brighter than their baseline fluorescence, so it seems odd that the CNMF-E is not catching them. Can you let me know if there is a standard-practice workaround for this issue? I can provide video if needed.

Thanks!

[pgunn]

Hello,
If you can share with us your data (please email me at pgunn@flatironinstitute.org with any google drive or dropbox share link; posting it here would let anyone else see it) I can see if I can find any set of parameters that might be able to find them.

--Pat