Error out of bounds while measure object intensity multiple channel module

I run it in Cellprofiler3.1.8 and 3.1.9 on Windows 10. Thank you for any help!


3_measure_cell_20200226.cpproj (1.4 MB) ROI002_ROI_002_probability_cell_mask_resize.tiff (588.2 KB) ROI002_ROI_002full_stack.ome.tif.zip (4.4 MB)

I’m not familiar with that module, and I wouldn’t be able to run your pipeline without it; it’s not in our standard plugins repository, for example. Can you tell me a bit more about where it came from and what it’s supposed to do?

It is a plugin I get from here. https://github.com/BodenmillerGroup/ImcPluginsCP/tree/master/plugins
It is used for multiple channel images, thank you.

Ah, then, I’ll cc in @votti .

Does it always fail, or just on a particular image set? If the latter, could it be corrupted?

I tried to re-install Cellprofiler and plugin or used different images, It always fails.


Thank you for helping!

Thanks for cc’ing me.

This module only works for multichannel color images.

Based on the error I think you have set the image in ‘NamesAndTypes’ as ‘Greyscale’ and not as ‘color’. Can you try setting the imagetype to ‘color’ in ‘namesandtypes’?

Also make sure the number of channels set in the measurement module matches the number of channels the image actually has.

Let me know if this works.

I am currently traveling for holidays without reliable internet, so allow some time for further answers.

Thank you for your prompt replying.
Yes, I set the mask image as “Greyscale” and the full stack image as “Color”, it doesn’t work.


I actually have 46 channels.

Hey,

The issue is that ome.tiffs are not not read as multicolor stack by cellprofiller. I usually pre-convert the stack into a stack that works with cellprofiller using the ‘ometiff2analysis’ function of our imctools package: https://github.com/BodenmillerGroup/imctools
See this example script how this can be achieved: https://nbviewer.jupyter.org/github/BodenmillerGroup/ImcSegmentationPipeline/blob/development/scripts/imc_preprocessing.ipynb

I hope this helps.

I will try it, thank you for your help, have a good trip.

Hi Votti,

I am using the “CorrectSpilloverApply” plugin, the error is " incompatible dimensions".
The spillover matrix dimension is 36mmx36mm, my images dimension are variable. How can I make them compatible? Thank you!

By the way, “out of bounds” error is also solved by Fiji-Cover to Hyperstack, adjust the channel to 46 and the slice to 1. Hope it can hope others.

Hi,
The spillover matrix has as dimensions the ‘number of channels’ measured. So 36x36 means that the images are expected to have 36 color channels. The xy dimensions do not matter, so this works for any image sizes as long as the images have 36 color channels.

Note that the order or the channels in the spillover matrix and your data must agree. I.e. the spillover image needs to be re-generated whenever you have images with different channels or different channel order (script https://github.com/BodenmillerGroup/cyTOFcompensation/blob/9867159e8cf4d975150029e0b5b60e25e6da1275/scripts/imc_adaptsm.md)

Does this make sense?

Yes, it works, thank you!