arivis Vision4D 3.2 Release Notes April 2020

Welcome to the arivis Vision4D 3.2 Release Notes - Version Date: 24 April, 2020 -

With a new clean Analysis Pipeline user interface users have a good overview over their analysis pipeline at any time. Paired with several improvements to existing Operations and the addition of completely new Operations, e.g. Seeded Region Growing Operation, users now have the tools to segment any object of any shape. Scaling via headless operation is also possible.
Now included is a set of new Sample Pipelines for common analysis tasks and detailed descriptions for an easy start.

Machine Learning Improvements

  • Work with colors
    • Machine Learning can now be trained on separate channels or for channel combinations. This allows color mixtures with co-labelling to be detected and quantified.
  • Import trained ilp files
    • We now support the import of pixel class ilastik ilp files, including 3D object features and color information.
  • Reuse training for several files
    • All training sets can be applied on other image sets which were not used for training one by one or in Batch Analysis.

New Analysis Operations

Flexible Seeded Region Growing 

  • Use information from different channels for better results
    • The Seeded Region Growing Operation uses two different channels to achieve more precise segmentation results. In a simple two-step approach, you can choose one channel for seeding and another one for growing. You can easily segment whole cell bodies based on positive nuclear staining to easily separate close cells.
  • New Watershed Operation for easy detection of non-roundish cells
    • The Blob Finder has a brother! This Watershed splits objects of any shape.
  • Combine Region Growing with Blob Finder or Watershed & VR
    • Create great results using any segmenter for seeding, using these as a starting point for Region Growing. Also, objects added into the data set via Virtual Reality can be used for Region Growing.
  • Easy splitting within the pipeline
    • Use the new Splitting Operation as a modifier of segmentation objects to automatically split them in a pipeline run.

See Overview for the different approaches to cover almost all possible segmentation tasks combining these tools efficiently. Watch Tutorials for more details on how to get executional segmentations results also for touching cells.

Analysis and Batch Analysis Improvements

Distances: centroid to surface measurements

  • Tune your Distance Measurements
    • Object distances can now be measured by surface-surface, centroid-centroid or surface-centroid basis. This allows for a custom selection of accuracy versus speed. If the particles to be measured against a big surface are small and spot-like, you can choose surface-centroid mode for high speed.
      In contrast, detail measurements with surface-surface distances give accurate distance measures of surfaces of larger objects.
  • Get your Distance Measurements faster
    • Several improvements in this Operation boosted speed for all measurement types.
  • Combine distances with Tracking and classification easily
    • The Distance Measurement Operation is an integral part of the arivis Analysis Pipeline and can be combined with other Operations such as object filtering or tracking, and directly group objects based on their proximity. Distance Measurement also allows you to easily identify classes of objects in certain distance regions inside or outside a cell or touching the surface. In addition, a Distance Map may be used for measurements and further processing.


  • Segments can have "children"
    • Individual segments can now have “child” segments. This removes the need to create groups for general parent/child relationships. Consequently, segments can have a count of associated objects based on distance or containment and can be tracked and further filtered in the pipeline. This supports easy and flexible workflows in a single pipeline run, such as “kiss-and-run” experiments with a combination of tracking and distance measurements or hierarchical relationships


  • This functionality is useful to combine multiple Analysis Operation outputs in a single step and is helpful in setting up user-specific statistical analysis.

Group by feature

  • Segments can now be grouped by feature and sorted into subclasses.

Batch output

  • In Batch Analysis, results of different image sets can be combined into one *.csv file. This is useful for data consisting of several image sets, such as multi-well plates, which all require the same analysis operation but need to be statistically analyzed altogether.

Magic Wand Improvements

  • More intuitive click & create behavior in 2D and 3D
  • Easy preview in the 4D viewer

Scalable Analysis - arivis Image Hub

  • Scale up your image analysis with efficient batch processing on a server
    • arivis Vision4D is now compatible with functions of the arivis Image Hub.

New Sample Pipelines and Pre-Installed Guides

  • New Sample Pipelines have been added and all sample analysis pipelines for typical image analysis tasks have been updated. New pipelines:
    • Detect cells or particles using seeded Region Growing
    • Detect small structures using Watershed
    • Compartmentalize cells or particles
  • Comprehensive descriptions of these pipelines are now part of the installation and can be accessed via the help menu