arivis Vision4D 3.4 Release Notes January 2021
Welcome to the arivis Vision4D 3.4 Release Notes - Version Date: 18 January, 2021
Release 3.4
Top Features
- New License Manager with a better overview of all licenses
- Machine Learning improvements (for multiple images sets)
- Easier to use and start of analysis with the new operation list, search options
- New Image and Object Math operations
- Python 3 Scripting support
4D Viewer
- Storyboard can export 4k movies (h264)
- improved sampling options in the preferences
- better surface quality for very small or thin structures
Analysis & Objects
- Redesigned operation list with short descriptions, categories, and search functionality
- New Image Math operation - do calculations between channels, planes, and timepoints
- New Object Math operation allows to subtract, merge or intersect object
- Improved and better-structured selection of input channels
- Newly created channels can be used in the same pipeline immediately
- Object Feature "Relative Volume" - volume percentage of the total volume
- Split objects by shape (w/o Analysis Pipeline)
- Much faster loading and sorting of objects in the objects table
- Speed of interpolation in the Draw Objects tool
- Basic support for high precision marker objects (created with arivis VisionVR)
Machine Learning Improvements
- Ability to define and use machine learning classes across several image sets
- Training for different pixel scaling
- for more speed
- more coarse textures detectable
- Redesigned feature dialog
- Better compatibility with the import of Ilastik trainings
- General speed improvements:
- Machine Learning operator now makes better use of available CPU cores (speed improvement for >10 CPU)
- Better compatibility for imported Ilastik Machine Learning trainings
Scripting Improvements
- Scripting in Vision4D (Script Editor) has been updated to Python 3
- NumPy is now part of the embedded Python 3 environment
- New Python integration with better support for debugging and external Python installations
- Script support for export and import of labeled images
- Run Deep Learning networks in arivis Vision4D:
- Python package and startup documentation supplied to run a prepared Tensor Flow library with V4D 3.4 via Python. This allows also non-experts in Python and Deep Learning to use established Deep Learning networks in facilities.
General
- The new License Manager allows to select a specific Vision4D license on startup
- Import of Imspector OBF/MSR data
- Support for the newest Lavision OME format
- Check whether an imported file should be discarded or kept when closing
- Color settings:
- Apply color settings to all image sets
- Perform Auto Color Range over all image sets
- Ability to reset color settings
