Color legend of 'Measure object intensity distribution'

cellprofiler

#1

Dear developer:

No one can deny Cellprofiler is much easier to handle and gives better result over other software on image analysis. The new module '‘Measure object intensity distribution’ is really cool! I love this kind of heatmap very much. It gives very straightforward result of the distribution. I do think it is perfect

The current problem is, if I want to save the heatmap generated from ‘Measure object intensity distribution’ module, the color legend is missing. Am I doing something wrong? Or it is a bug of the current software?

Hope to get your response.

Best
Dingcheng


#2

Glad it’s helping you! And you’ve described your question so clearly.

I suspect that is just not an option to save the legend - we use a single function to save the output of most modules and the majority do not have a legend so we just don’t have that feature. As a workaround, you could take a screenshot of the legend for an individual case, though obviously that does not help you if you are wanting to save lots of images’ output.

We used to have the option to save the ENTIRE figure window (including all the contents therein) but I believe this went away, perhaps someone else can confirm that?


#3

Hi:

Thank you for your reply.

Maybe in the next version of CP, the technician can improve this. I hope so.


#4

It’s actually non-trivial to do this, for boring reasons (though I can explain if need be). And yes @Anne_Carpenter, that functionality to save whole display windows did exist in 1.X but not since the switch to python in 2.0.

Is the reason you need the color bars so that you can compare between multiple images? Because one thing I think we could maybe do that would be much less difficult to implement would be to add using an an absolute range for the colormap rather than a relative colormap for each image- then you could compare between images and just use the colormap reference. Would that be a suitable workaround?


#5

Hi bcimini:

Thank you. That should be one of the solution. In my case, I use the ‘Fraction at distance’ for the distribution. So I think all images are already using the same absolute range.

If it is impossible to add the color bar for each image, a manual operation is required at last.