I am getting started with CODEX. It has its own workflow using an ImageJ plugin, but it’s relatively limited. I’ve heard that StarDist is a good alternative segmentation algorithm. It exists as a Python library and Fiji plugin. Has anyone had any luck with either option? I was hoping to find some example protocols, but couldn’t. Essentially, I am trying to figure out the proper inputs. Do I need to pre-process the CODEX data in a specific way? Can I use the pre-trained models or do I need to generate my own?
Are you opening the CODEX images in QuPath, since you have that tag? Or is the problem that you cannot open the images in QuPath?
If the images are being read in correctly, that should be all you need - otherwise they are basic immunofluorescence images. Just with a lot more channels. You may need to make sure the metadata is being read in correctly for the pixel size.
I have not yet tried to open them in QuPath, but I saw that was an option. I wasn’t sure if there are some additional pre-processing steps I should perform.