Trainable weka segmentation fails

Hi all,
I have got a video with 60 time-points; there are 25 frames (2D image) in each time points.
I tried to segment the whole video it by using “Trainable weka segmentation” but it fails. Do I need to segment it for each time-point? Is there any script to do that?

Thank you very much for your support.


Hello @hello_fiji,

The plugin might run out of memory depending on the size of your frames, the number of features you use and your RAM memory. Can you tell us those numbers?

That being said, you could start by 1) segmenting only one frame and 2) reducing the number of features in the Settings dialog and see how it goes.


Hi @ ignacio,
Thanks for your reply. I am using only five features (default features). I have got 6 GB RAM.
Since I have 58 frames (each with a stack of 23 images), I would like to know how do I create 58 stacks easily (I can do it manually using Fiji: Image --> Stacks —> Tools —> Sub stack).

Once I have 58 stacks, and if I create a classifier from one stack, can I apply the classifier automatically (kind of a batch processing) to all the stacks?

Thanks again.

Hello @hello_fiji,

Can you train on a single image and try to segment each stack?

Regarding how to create the stacks, you can make a short macro out of the same command you mentioned.

Yes, you can either create a script and call it from command line, or directly use the Apply Classifier option in the GUI.

1 Like