Version 1.73.0

Version 1.73.0

Trainable Weka Segmentation nodes

JIPipe nodes and data types encapsulating the Trainable Weka Segmentation were added, making it possible to utilize these without the Macro node.

  • New data type “Weka model” (saves classifier, data, and metadata of the segmenter)
  • New node “Import Weka model” (allows to import existing models)
  • New node “Train Weka model from ROI (2D)”
  • New node “Train Weka model from ROI (3D)”
  • New node “Train Weka model from mask (2D)”
  • New node “Train Weka model from labels (2D)”
  • New node “Weka classifier 2D”
  • New node “Weka classifier 3D”

Using the Weka segmentation nodes

A small example showcasing how to utilize the Trainable Weka Segmentation nodes.

Skeletons

It was requested to include “Analyze skeleton 2D/3D” as node, so the user did not have to utilize the Macro node.

  • New node “Analyze skeleton 2D/3D” (tip: Use the “Outputs” parameter section to enable more outputs)
  • New node ” Morphological skeletonize 3D”

Using the Analyze skeleton 2D/3D nodes

A small example showcasing how to utilize the Analyze skeleton 2D/3D node.

ROIs and labels

New nodes were added that simplify the processing of ROI lists and label images

  • New node “ROI to labels (Expression)” (the label value is assigned via a custom expression)
  • New node “ROI to labels (by name)” (the label value is assigned by the name of the ROI)
  • New node “Labels to ROI” (based on https://labelstorois.github.io)
  • New node “Color ROI by name” (tip: if you want to color by statistics, there is “Color ROI by statistics”)

Conversion between ROIs and labels

A small example showcasing nodes for conversion between labels and ROIs and vice versa.