Want to generate Max projection and use in same pipeline downstream

I have image stacks with 4 different channels and I’m trying to analyze these images at the level of individual stack for two channels and in max projected image in other two channels. I need to use these Max projected images for analysis of other two channels.
the problem is Max projected image comes in last cycle so I have to make the max images in separate pipelines but now when I load these images with my other stack images they do not align well as there are more stack image than max intensity images. there is mismatch of images happening (screenshot attached [Image name in single row are different

To load special “single” images like the MAX projections in CellProfiler, you need to use the “Add single image” button at the bottom of NamesAndTypes, like on the right here:

This is also useful for loading illumination correction functions.

Hi David,
Thanks for replying but I don’t have one single image like Illum function image. Here I have 70 Max projected images from 70 stack frames. that too 70 max projected images X 4 channels.

I’m also trying to make max projection in same pipeline but it gives following error-

Traceback (most recent call last):
File “cellprofiler\pipeline.pyc”, line 1934, in run_image_set
File “cellprofiler\modules\makeprojection.pyc”, line 164, in run
File “cellprofiler\modules\makeprojection.pyc”, line 389, in accumulate_image
ValueError: shape mismatch: objects cannot be broadcast to a single shape

I’m not sure where the mismatch occurs, because frankly, I don’t understand what images you want processed together. By definition, there are fewer MAX projected images than raw images, so how should they match up with your original images? Can you give an example of the image names?

If you want to compare, say, a single slice of your stack to the Max projected image, then the numbers would match. You would likely need to use the Metadata module to extract the matching parts of the file/folder names between these two input types and then match them in NamesAndTypes.