ATTENTION USERS: Help us with usability improvements!



Thanks for giving the new interface a try! Can you tell me what you are doing when either error msg occurs?

It seems like you were at least able to attempt to run a pipeline. Could you post the pipeline and a sample image set?



Hi Mark,

I just click: “continue processing”. Anyways, this does not help.
I was now able though to run the pipeline after removing “Illumination correction” and “rescaling intensities”. The pipeline runs much faster this way (so it really seems to use more than one core). However, when I tested the pipeline with 2 pictures and compared the results with the same pipeline run in Cellprofiler 2.0, the results of the analysis differed. I also cannot use any graphical output. So I have no chance to visually see whats going on. Is the graphical output not ready yet?

I will upload my pipeline (with and without illumination correction) and 2 sample images. It would be great if you could provide some quick help. Thanks a lot!!

Edit: Just tried it on my laptop again. Here it seems to work. However, I dont have any visual feedback…not even the illumination correction status is shown. (726 KB)


It makes sense that they would differ since you are no longer rescaling the image but still using a manual threshold in IdentifyPrimaryObjects.

I’m not sure what you mean by “I don’t have any visual feedback”. Are you saying that as the pipeline runs, no module display windows are shown? If this is the case, is there an error that’s written to the terminal window but not halting the program?



A much-belated response: The ability to enable/disable modules was been implemented with CellProfiler 2.1.0, by clicking the green checkmark next to the module name.


hi all,

I think maybe there’s a problem somewhere because I can’t run two or more cell profiler projects simultaneously. If I do, I get the error attached. I am I doing something wrong somewhere? Thank you as I look forward to hearing from you guys.



Hi Raphael,
Are you encountering this error when you run the same project with multiple instances of CP, or different projects?


[quote]How can we make IdentifyPrimaryObjects easier to use?
For example:
What current features/settings do you find easy to use? Difficult to use?

The ONE thing that made me move to CellProfile, besides the High throughput design, is the little tutorials everywhere with the “?” and the inclusions of references. I really like the examples and suggestions inside each algorithm explanation. Making an effort to make these explanations of the algorithm something more contextualized (with examples, situations that can be better applied, possible bad consequences) can be of huge help to lay people in image processing, and makes the program more self-teachable.

ImageJ. But I prefer Cellprofiler now. however there’s one thing that is way easier to use in ImageJ: Manipulating images. EditingObjects Manually and such stuff are pretty easy there. But in CellProfiler, maybe because of the plataform (I don’t really know), each drawing takes a lot of processing. I usually use ImageJ to draw my masking images in binary format and load them in CellProfiler. It was an option around the difficulty of drawing the ROI for example, in the image.

The example of clicking in the object and get its features would be awesome. Could make the “morphology-image metadata” relation that lay people (in img processing) find difficult a lot easier. Another suggestion is in the first question.

For CellProfiler, I think this is dangerous. It may take a person the drive to explore more, since there’s a simple interface, or worse, deprive them of the possibilities they can have to process their images. I think that a great effort in a more exemplified or detailed instruction (in the ?s) could help.
Also, If some common processing (like the basic pipelines) could be described step-by-step explaining what they do in each (in general lines) and why they are necessary or desired, could help
When I started on CP, my biggest question was “What do I need to do with my Image in each step?” Maybe this “Simple version” could be a sub-interface, chart-flow based, asking questions and presenting the more appropriate options depending on their type of Image and need and etc. Each step would present a number of options depending the users need. This can take some options off the table, but would guide people initiating in a sense of teaching them what they need to (or can) do with their images.

Hope to have helped.


Hi Mark,
Some suggestions:

1- Here’s a possibly useful nuclei segmentation algorithm:
Qi, J. 2014. Dense nuclei segmentation based on graph cut and convexity–concavity analysis. J. Microsc. 253:42-53. recently mentioned in Kurt Thorn’s blog. While I’m not convinced the graph-cut segmentation would do better than what CP can do, I think the convexity–concavity criterion for nuclei splitting has some merits. I run it in ImageJ (from CP) and it helps resolving some annoying little murine fibroblasts that pack together.

2- User-defined variables (perhaps loaded from a csv file) might be useful when mixing cell types that require different parameters (e.g. smoothing radius, threshold correction factor). Effectively, enabling the right-click insertion of numerical metadata in fields such as “Threshold smoothing scale” or “Threshold correction factor”. In my case, with murine and human nuclei it’s not easy to find compromise parameters good for both, so I’d very much like the option of a csv file where I put all the processing parameters for each metadata set. And that would keep an easily accessible record of the parameters used in each run.

3- In relation to loading metadata, in the option “Import from file” an easier way of specifying the folder (that is, not having to browse to a file for every new set of images) might help (either constructing the path from metadata from other extractions in the module, or using the default import folder, or looking in the dragged tree of files).

Hope it helps
Thanks very much for your work


Just a thought after clicking on the wrong thing a few times:

Different icons for project and pipeline files?


I admit to making the same mistake myself; I’ve opened an issue on it here: … ssues/1396


More Annotation Options. I know each module has a place for notes, but what i would really like is to be able to add annotation to or near the module name (from what i understand the modules are automatically named). It would help me navigate the pipeline a bit easier during auditing, and tweaking.


You are not the only one who wants this! I’ve added it to our Github issues list here so you can see whether there is any activity on this:



Regarding the loading of images and general workflow when working with multi-well plates. I was wondering if you are thinking about implementing a plate map mode in order to make it more dynamic? For instance, after loading the images and the metadata, a plate map would apear on the left side of the general GUI window, one could then click on a well to select an image to set up the analysis, and so on.

Ideally, one would also have the option to chose what channel to display, and so on.

This would make it so much more practical and useful overall.

Software like the MetaXpress or InCell Analyzer have it, and it makes it a lot easier to set up the analysis for high-throughput.



Feature Requests: Image Sets Options

Hi Jordi,

The short answer is no, at the moment it’s not under consideration, but I’m also not certain how this would be used or what this would take the place of. Are you thinking of something that would be used during/instead of NamesAndTypes, something that would be available the whole time you’re in test mode, etc?

If you can give me more specifics about what you mean and where this would be used, I can let you know how feasible the given changes are or if there are already good workarounds in the software.

Thanks for the feedback! We are always glad to hear what people would find helpful!



Thanks for the answer.

In the way that the plate layout is used in other software is to basically navigate through the images of the plate in an easy, visual and direct manner.

The plate layout would be always present, this would allow for instance, with only one click, to compare if the analysis settings work for both positive and negative controls, and eventually for a given sample.

The way that I see it is that once the images are loaded, the metadata has been retrieved (from a file, for example), and the names and types and groups have been processed, the plate layout would appear and stay there, as I said, to allow straight forward and direct navigation to the images.

Another feature would be to indicate an ongoing analysis, meaning, the wells would change color as the analysis is finished, just to indicate the progress.

Ideally there would be a small drop down menu next to the plate layout that would allow to chose the different fields (if there is more than one) and the different channels for visualisation and for checking the current analysis set up.

I hope that this helps to more or less explain the plate layout functionality. I can certainly see a big advantage having such a feature. We use it in other image analysis softwares and it’s great.




Hi Jordi,

I think that’s a really interesting idea, and I’ve made it an official request on our GitHub. All of the components basically exist already (we do have a Plate viewer, which is accessible under Data Tools-> Plate viewer; you can choose which well’s images to use in Test Mode by going to Test->Choose Image Set (you can even sort by row, column, plate name, or any other metadata (Pos/neg ctl, drug dose, etc) you’ve passed to that point); there is a progress bar during analysis), so it seems like it ought to be possible.

Our software engineers are really busy though getting CellProfiler 3 ready, implementing the new 3D functionality, fixing bugs, etc, so my prediction (which is just a prediction, not a certainty) is that user interface changes are likely going to take some time; once we’ve added the new features there will be more time for streamlining the interface like how you’ve described. You can check on the link though for the status of the request. Thanks for the really clear feedback!



Thanks a lot for making it an official request, I think it would help.

Thanks a lot as well for the tips, so manye years using CellProfiler and discovering functions every day :wink: The plate viewer doesn’t work for me for some reason. The option of choosing the image set is really useful and I was not aware of it :grin:

Looking forward to CellProfiler 3 and the new implementations!