I am very much interested in the dissection of cells from pictures with nuclei stain (hoechst) and cytoplasm stain (phalloidin). I tried in Knime several methods to dissect the cytoplasm (voronoi using nuclei as seeds or watershed with the same seeds) however the resulting segmentations were not verry good.
However previsously I was working with propagation method in CellProfiler and this method provided solid results, here is a link to the publication:
My question is the following: is this tool available in Knime (if yes I could not find it) or is it something that might be similar?
Thank you very much in advance for the help of the Knime community,
unfortunately this method is not available in KNIME Image Processing, ...yet. I have put it on our list for algorithmswe want to integrate. thanks for the link!
Anyway, you have two options:
* If you want you can still use this method and do your measurements in CellProfiler and propagate the results back to KNIME. We have just released a CellProfiler integration for that (see: https://tech.knime.org/community/cellprofiler)
* The Voronoi-Segmentation is the closest we have to the method which is described in the attached paper. Maybe tweaking the parameters helps there (e.g. setting a threshold or preprocessing the images with some filter).
Sorry that I don't have better news!
Thank you very much for answering so fast. By testing it seems as you said that the voronoi-segmentation gives the most similar results so far.
Do you have an idea of when the propagation method might be implemented in Knime?
I can't tell you a date, yet. But it's of course an algorithm we want to make available in KNIME. Do you know about any ImageJ plugin implementing this functionality? Then an integration would be pretty simple and I could do it immediately.
Have a nice week,
I did not find an ImageJ plugin for the propagation method yet. If I do so I will not forget to post it on the forum.