I am working with multi-dimensional confocal images datasets: X,Y, C and T. The metadata is read correctly with ImageJ, I can see the time information (30 time points, at 10 seconds intervals) and the calibrated pixels (0.88 microns/pixel). The Knime image reader sees five dimensions in my dataset, X,Y,C,Z and T, and it recognizes their sizes correctly: 100 pixels for X and Y, 4 channels, and 30 time points. But I need to work in pixels and seconds. I tried the “Set image metadata” and “Image properties” nodes for that purpose, without success. I don’t seem to able be to manipulate the metadata to output microns (instead of pixels) and seconds (instead time point 1 to 30). Anybody resolved this? Thanks!
Is it possible you manually convert your quantities afterwards based on the metadata? You could certainly follow that approach with KNIME using the Image Properties node (select the Calibration feature). That will give you information on the distance between your samples in each dimension, although I am not entirely sure in which unit.
Maybe @dietzc can comment on that one as well?
Thanks for your reply. I already tried the Image properties node, and it is not providing the time interval. I also tried what you suggest, that is do the conversion afterwards based on the metadata. I works OK, but in the long term, that is not practical since there will be in the future a lot of datasets coming from different microscope, hence slightly different metadata. I was hoping to set microns and seconds at the beginning such that I can simply apply the rest of my workflow to all my different datasets. I was hoping to achieve my analysis faster with Knime than with ImageJ.
I’m following up on the chat we were having on the KNIME forum. Sorry for the cross posting, but I wasn’t sure where @jlacoste would get the quickest answer, so I recommended both forums to her.
We can post an example, of the image and workflow @jlacoste is working on in the next couple of days. In the past I’ve tried to post workflows by exporting them, but several times they have ended up too big (even if I don’t export the data).
Also we would prefer a colaboritive workspace. So we can basically point people (including each other) to where our latest work is, as we have questions or just want to share.
We have already been sharing workflows on google drive, so we might just clean up that space post a link. If anyone else has other suggestions on the best ways to share KNIME workflows, let us know.
One way to share workflows is to create a workflow which we could upload to our Example Server (if the use-case is general enough). Otherwise, I’ve seen people sharing workflows via GitHub & Co or just upload them here in the forum. I’m wondering why the workflow became too big without data. Do you have an example for that? Usually it shouldn’t be bigger than max a couple of MB.
Anyway, we want to make sharing workflows easier in the future and are thinking and working on some ideas, however, I don’t have a real timeline, yet.
Let me know if I can further help