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
Due to the small size of these datasets, and due to the way cells are selected, most of the results we are seeing are really about cells from the mean image, which as a majority do not have activity. I don't care much about these cells, and no one should, because they are overwhelmingly neuropil contaminated. It is perfectly possible that an algorithm detects a lot of these silent cells and does very well by your metrics, while doing very poorly on the 10% of cells that actually matter: the active cells.
I would suggest labelling every cell in your current ground truth as active or inactive, and then also running all the benchmarks on the active subset only. There could be a switch at the top of the website to flip to "active cells only". The definition of active should definitely subtract off neuropil from each ROI, before quantifying something about the variance of the trace, perhaps relative to very high-frequency content of that trace.
The text was updated successfully, but these errors were encountered:
This would be particularly relevant for the 00 series, where the nuclear marker is in all cells, but gcamp is not even present in a large fraction of the cells.
Due to the small size of these datasets, and due to the way cells are selected, most of the results we are seeing are really about cells from the mean image, which as a majority do not have activity. I don't care much about these cells, and no one should, because they are overwhelmingly neuropil contaminated. It is perfectly possible that an algorithm detects a lot of these silent cells and does very well by your metrics, while doing very poorly on the 10% of cells that actually matter: the active cells.
I would suggest labelling every cell in your current ground truth as active or inactive, and then also running all the benchmarks on the active subset only. There could be a switch at the top of the website to flip to "active cells only". The definition of active should definitely subtract off neuropil from each ROI, before quantifying something about the variance of the trace, perhaps relative to very high-frequency content of that trace.
The text was updated successfully, but these errors were encountered: