Batch processing with ilastik. Cannot load stack of images

I have trained a model in ilastik (1.3.2. on Ubuntu 16.04) on one of my (microCT scanned lung) datasets.
Now I’d like to apply the trained model to the 11 other datasets.

The dataset I trained ilastik on is generated from a stack of images (approximately 700 images with 700x700 px), which I loaded as a stack.

I cannot seem to find a way to apply ‘Batch processing’ to the other 11 stacks.
The only option I have is ‘Select Raw Data Files’, which seems to expect one file per sample and not a stack of files per sample.
Do I need to convert the other stacks to H5 files prior to loading them in ilastik?

1 Like

Hi @habi,

unfortunately that’s true :frowning:
There are two ways around it:

  1. as you have mentioned yourself, converting the stack to a single h5 file will do the trick.
  2. In the headless mode you can specify stack input

let us know if you need any help setting up the command line for the headless mode, or if you are unsure how to convert the data.


Thanks for the pointer, I should have read the help more thoroughly!
Going from the saved project to applying to two of the 11 others seems to be just

/$path_to_ilastik/ --headless --project=$filename_from_training.ilp "dataset02/rec_crop0.85_skipped50lowest_slices/dataset02*.png" "dataset03/rec_crop0.85_skipped50lowest_slices/dataset03*.png"

(slightly redacted for “privacy” reasons).
The process will take a while, if I don’t write back, everything workded as I think it should have…

1 Like