CellProfiler 2.1.1 release!


#1

The CellProfiler team is pleased to announce the release of CellProfiler 2.1.1! Download it from cellprofiler.org.

In this minor patch update, we have dealt with a number of bugs, mostly reported by our faithful users :smiley: Also, we’ve improved color image stacks handling, improved performance for image stacks in general, and improved compositing of overlays on images. More details are in the release notes.

This update should be considered non-essential but really, why wouldn’t you want to update? :smiley:

Happy profiling!
The CellProfiler team


#2

Cell Profiler crashing when I try to use Names and Types following update


#3

Hi mfennell,

Is there an error log you can report? Or does the whole application crash?
Are you extracting metadata first? If so, in Metadata, have you clicked “update Metadata” and inspected the table for missing metadata?

Thanks,
David


#4

Dear davidlogan,

Cellprofiler stopped working for me well when I used the input module.
I’m working with .czi files and was happy to discover that cellprofiler can now read this format. So, I followed the instrustions of “how to read in .czi image stacks”. In particular, each image consists of 2 slices (channels).
It worked fine for a test selection of 10 images. However, I can not get cellprofiler analyzing a larger number of images (165). The Metadata extraction from the file header takes ages and at the end it does not only detect both channels per image. As you were asking the author of the previous post - yes, the channel metadata is missing. I assume that this module error cause conflicts further downline so that the pipeline is interrupted.
What can I do to improve the metadata extraction?

Thanks and best,
Richard

Error report:
Traceback (most recent call last):
File “/Applications/CellProfiler.app/Contents/Resources/lib/python2.7/cellprofiler/pipeline.py”, line 2094, in prepare_run
File “/Applications/CellProfiler.app/Contents/Resources/lib/python2.7/cellprofiler/modules/metadata.py”, line 725, in prepare_run
File “javabridge.pyx”, line 1404, in javabridge.JB_Env.new_string (javabridge.c:13827)
MemoryError: Failed to allocate string


#5

Hi,

I don’t know if this is the place to report this, but I think there is a bug in the ImageMath module. I had a pipeline written in 2.1.0 that subtracted 2 binary images (no manipulations to them. just a simple subtract image 2 from image 1. It does not work in 2.1.1.

The results look like this:
(image 1) - (image 2) = new image
1 - 0 = 1
1 - 1 = 0
0 - 1 = 1 (this should be 0. Yes, I have turned on “set values less than 0 equal to 0”)

I also tried swapping the subtraction order of the 2 binary images and I get the same results (if it were working correctly, I should see an inversion of what I want)

Perhaps I am using something wrong or there were changes to the module that I have not accounted for? I did not see any new options or settings in the module since 2.1.0.

I have reinstalled 2.1.0 and it is working fine.

Thanks in advance.


#6

Hi scottkg,

I suspect that this is the same issue which we discovered recently (Github #1171). Glad that you have a workaround with 2.1.0, and we will try and remember to post back here when it is fixed.

Thanks,
David


#7

Hi richardw,

Sorry for the delay (FYI it’s more likely to get our attention if you start a new thread rather than add on to an existing one)
Your error is low memory. Unfortunately on a Mac, we are limited to 32-bit memory addresses. It’s annoying to say, but if you can run this on Windows, try that and you may get get it to work (assuming you have >4GB RAM to take advantage of the 64-bits). Alternatively, have you tried specifying the metadata from the file and/or folder info, rather than extracting from the header? I’m not sure if there is enough info in those from a czi file to specify each metadata you might want or need.

David


#8

Dear all,

I installed cell profiler which complains about a missing module “ilastik_pixel_classification” which is listed in cellprofiler/modules/init.py but I found no ilastik_pixel_classification.py in the same folder. Did I forget to install something ? I checked out the commit tagged as 2.1.1rc4.

Regards, Uwe


#9

Hi Uwe,

Nope, you didn’t miss anything – this is a harmless warning. It is a CellProfiler module kindly contributed by the ilastik folks for their 1.0 release. We have not validated it yet however and so only support the 0.5/0.6 versions of ilastik. In any case, it should not affect CellProfiler.

Cheers,
David


#10

@scottkg,

Github issue #1171 has been fixed which we hope addresses your ImageMath issue. You can try a bleeding-edge trunk build of CellProfiler here: github.com/CellProfiler/CellPro … TrunkBuild

David