A collection of scripts for simple computing in cluster
Input: Various image files
Output: processed images / stacks
Use of ImageJ in cluster is preferred with large data sets but needs “headless” processing, which means that the computation should be done without graphic user interface (GUI). As ImageJ is tightly coupled with GUI, running a script developed in desktop environment often fails when it is executed in cluster, where no display is available.
For this reason, scripts should be written specifically for use in cluster. This page lists some of such scripts, which could also be used as a template for different types of processing.
ImageJ, some of them require Fiji where indicated.
To run the script in cluster, the script must be placed somewhere accessible in the network. To tun a scrip, following command is an example.
ImageJ -script <full path to the script file> [arguments]
Following is the list of headless scripts. These scripts are all located under https://github.com/miura/ijmacros/tree/master/headless
Opens a image data set from 3D-time series LIF file (Leica format) and outputs Max-Intensity Z-projected 2D time series.
Opens a image data set from 3D-time series LIF file (Leica format) and outputs Sum-of-Intensity Z-projected 2D time series.
Opens a image data set from 3D-time TIFF series and outputs Max-Intensity Z-projected 2D-time TIFF series.
‘Correct 3D Drift’ is a Jython script preintalled in Fiji. This script allows to run this command in headless environment. Requires Fiji.
Using a trained model generated on desktop (this requires GUI to provide manual annotation of classes), this script does trainableWeka segmentation. Requires Fiji.