Allen Cell Discussion Forum

No module named 'aicssegmentation'

Hi Mr/Ms

I am new into using the Allen Cell Structure, I had problems during the installation of “pip install -e .[all]”.
I tried to dig in what could be the problem, also tried to find a solution.

I ran the command pip install pandas and it was successful.
I also ran up to a suggestion of running this command “git pull” thinking maybe the problem was with windows setup, but the output was “Already up to date”.
The error comes when running scikit-image. I would appreciate your help.

Sorry I meant “pip install scikit-image”, was the successful.

Can you post the error message so that i could better understand the problem?


I have managed to fix the error which cost me a lot of time, I thought python 3.8 would work after rereading I saw that the tools only work for python 3.6, so I had to create a new environment on my anaconda and set it to 3.6.

I don’t know if you could help with this new error I am getting, the images I am using in my datasets they have got 5 CHANNELS (0,1,2,3,4,5) and I want to reduce to 4 CHANNELS, do you know a way of converting the number of CHANNELS in an image, because these images with 5 CHANNELS are giving me error when I try to run the code.

Thank you in advance.

hello, @FutureCScientist, I would be more than happy to help. Again, if possible, please copy and paste the error message you see, or even screenshot, so that I can understand what is wrong.

About multi-channel images, we are using aicsimageio to manage image reading and writing, you can find more details here:


(1, 1, 1, 1944, 2592, 3).

These are numbers OF CHANNELS in a picture,
Then below is the error I am getting, I think it might be because of the number of channels in a picture.

I will be looking into the aicsimageio.

Thank you.

The extension of images i’m using is jpg.

If i understand correctly, your image is 2D RGB image. Currently, the demo was made with 3D image segmentation as the primary target. We don’t have demo for 2D data yet. However, a lot of functions are compatible with 2D data. We may include bette documentation and demo for 2D data in the future.