Questions on AllenCellModeling/pytorch_fnet data set

I work on the data set from the Allen Insitute, used in this work, and which can be found here.

In each picture you get 7 channels, with (for example) the following names: CMDRP, CMDRP, EGFP, EGFP, H3342, H3342, Bright. It is basically tow channels for each fluorophore, with the first one always looking quite dark and empty.

How should they be treated? Is the first one of each to be considered as background, to be subtracted from the second one?

those black images are dummy channels, which can be ignored.

Jianxu

Perfect, thanks for your answer!