We'd need to retain the timepoint label for each bead ROI since that's essential to know for which timepoint it's to be used for drift correction (shifting or accuracy analysis), but we could save some filesystem overhead if we combine by FOV (which we'd still need to keep separate -- #75 ). For example, with 61 timepoints and 30 FOVs, we create 30 FOV 61 timepoints/FOV 2 purposes/timepoint * 1 file/purpose = 3660 files. We could reduce this by a factor of number of timepoints T = 61 to just 60 files.
We'd need to retain the timepoint label for each bead ROI since that's essential to know for which timepoint it's to be used for drift correction (shifting or accuracy analysis), but we could save some filesystem overhead if we combine by FOV (which we'd still need to keep separate -- #75 ). For example, with 61 timepoints and 30 FOVs, we create 30 FOV 61 timepoints/FOV 2 purposes/timepoint * 1 file/purpose = 3660 files. We could reduce this by a factor of number of timepoints T = 61 to just 60 files.