Error Help

I have been getting the error “zero-size array to min.reduce”. I’m not quite sure why this is happening or how to fix it. Can anyone give me an idea of how to correct this error?

Hi,

It’s hard to tell given just the this single line of the error message. Does this happen for all images? In what module? If you attach your pipeline (.cp file) and the image that is failing, we can help.
Also,

  • What CP version?
  • PC or Mac?

Thanks,
David

I am working from a Mac with cell profiler 2.0, i’ve attached my pipeline and picture (which is 1 picture split into 3 channels). The error shows up at the second to last module “measure object intensity”. I have a bunch of pictures and i think only one of them game me a read out but im not quite sure what makes that one picture different.

Thanks for your help!






p53MDM2.cp (11 KB)

Hi,

I am not sure what was causing the error, but I did reproduce it. It may have been that you had some measurements that you named with space and dash characters in them. We recommend that you stick to alphanumeric and underscore charaters.

In any case, I am attaching a much simplified pipeline. This does what you were doing, but using a couple IdentifyTertiaryObjects and a single MeasureObjectIntensity module to measure at least as many compartments as you had defined. I also added a couple comments in the Module Notes for some modules, reproduced here:

  • LoadImages: File selection method ‘Order’ is potentially dangerous, since we cannot guarantee the order and if you change OSes it might be different. I suggest using ‘Exact Match’ if at all possible and key off of “blue”, “green”, and “red”.
  • IdentifySecondaryObject: This didn’t seem to be segmenting properly, as far as I can tell. I switched the thresholding method from Propagation to ‘Distance - B’ and it seems much more reasonable to me.
  • ExportToSpreadsheet: Please re-choose which measurements you would like to export.

Hope this helps!
David
p53MDM2_DLogan.cp (10.6 KB)

Thank you very much! I definitely think this will help. I’ve started using 16 bitt images so I’ve had to tweek a few things but everything should work out!