Hi,
I’ve tried to run StarDist to create 3D segmentations in Knime using the Python integration. It works but until now I load the 3D input image stack in the python code from harddisk with the respective filepath transfered as flow variable. I also save the StarDist prediction to harddisk in the python code. I have seen some examples where 2D images were directly transfered from the input table to python and newly generated images were also directly transfered to the Knime output table. I wonder if this also works for 3D images and if there would be some small example workflow on how to set this up.
Hi Christian,
The new Python Script node does not yet support reading and outputting image cells from the Image Processing Extension. We are working on supporting n-dimensional images in this node. Stay tuned for updates.
With the Python Script (legacy) node it is still possible to read and output images from the Image Processing Extension with the community extension “KNIME Image Processing - Python Extensions” (The README contains an example of how to use it). This also supports 3D images.