Export objects separately

cellprofiler

#1

Problem 1: I’d like to load 500 different images, perform area and intensity measurements on objects chosen from the image, then export the data from each image to a separate file.
I can think of two potential ways of doing this:

  1. Load them sequentially but as individual files rather than as cycles, i.e., start a new Pipeline for each image in some automated way.
  2. Load them as usual (as if each image was in a different cycle), but then somehow save the object info. separately.
    Is there a way to do either of these (or any other way to accomplish the same goal)?

Problem 2: Similar to problem 1 except that there are five hundred groups of files, each with ten images representing ten time points of the same microscopic field.
For each group of files, I already know how to measure intensities for objects that are represented in all ten time points.
But I want my object output to be arranged in 500 separate files, each containing info. on the ten timepoints.
I again need some way to sequentially load sequentially in an automated fashion or save object info. separately.

Many thanks in advance,
ibv


#2

Hi,

In CellProfiler right now, there is no easy way to do this- the exported Excel file either is for Image, with mean & std dev measurements written one row per image (so 500 rows in your case), or for object, with the individual object measurements written one row per object (total object number in your case). You could utilize RunMultiplePipelines to accomplish the first way you suggest of obtaining your goal: each pipeline would be one cycle long, would have one Image Excel file written, so you’d get 500 files.

Similarly, you could accomplish task #2 in the same way- run 500 pipelines on the 500 groups of ten images.

~kate


#3

Kate has basically the right of it on this topic. We’re working on adding the ability to creating image groups so a user can treat subsets of the full image set as independent entities and export measurements accordingly, but that won’t be ready for some time from now.
-Mark


#4

Thanks Kate and Mark -
RunMultiplePipelines will do it in theory (automating the process), though I have some questions about it:

Running RunMultiplePipelines fixes both the input image and output folders as the folder containing the Pipelines. Is there a way around this (i.e., allowing input and output folders to be different than the folder containing the Pipelines?)

More importantly,
I can easily copy the Pipeline .mat files in Windows, making 500 of them.
But then I would have to define separate input images and output files for each of the 500 Pipelines. The only way I know how to do this is to set each Pipeline with their own input image file names and output data names. For 500 files, this would take a lot of time (and I would have to change all 500 files each time I want to change a setting!).
Is there a simpler way?
Thanks,
ibv