moldia / iss_preprocessing Goto Github PK

The function leica_OME_tiff has a line which hard-codes the shape of the OME tiffs to 5 channels (4+DAPI), see below.

stacked = np.empty((5, 2048, 2048))

This makes it impossible to work with images acquired with more or fewer channels than the "Nilsson standard" way. I realised this while trying to work with 5+1 images.

We should come up with a flexible way to specify this (ie, a variable defaulted to 5 unless otherwise specified?)

This condition of the leica_mipping function leads to missing tiles if the number of channel is not the "Nilsson standard" (4+DAPI).

if len(strings_with_substring) < 5

When mipping images with 6 channels, the output was missing every 6th tile. I've made a dirty fix for myself but I think this needs to be soft-coded (ie, user specified input?).

Christina brought this up, I think there's a problem with the loop through the m dimension of the czi files.
It's probably just sufficient to remove that loop and only loop for S.
Checking now (17 feb 23)