I am currently working on implementing cycling IHC multiplexing in our lab by using AEC as chromogene and stripping off the abs after scanning to re-stain the tissue with a new ab subsequently. I would like to co-register then the resulting images (still working on this) and like to know as what kind of image file I should save them to retain the multiple channels and analyze the multiplex via QuPath. I guess I would have to transform beforehand each AEC signal into a false colour, right?
OME-TIFF sounds like it would be best - this could also be a pyramidal image.
I have done a little bit of work on dynamically applying color deconvolution to multiple images and concatenating them along the channel dimension as the pixels are being requested, to avoid the need to actually write a new file… but this only supports affine transforms. If you need a more complex alignment then you’d likely still need to write a new file - or, potentially, a super-fast method of applying the transform.
(The hardest part of this for me currently is figuring out how to represent the required images & transforms within a QuPath project so that they can be saved & reloaded… which is the main reason it has not emerged as an actual feature in the software yet. For now this remains one of many unfinished projects and I’m not sure when I’ll have time to revisit it… so I’d recommend trying the OME-TIFF approach for now.)