arivis Vision4D 3.5.1 Release Notes November 2021

Analysis:

• Improved: error handling for missing input channels
• New: ability to match missing input channels
• New: more accurate mesh-based calculation methods for surface area and sphericity (using Marching Cubes)
• New: more accurate mesh-based calculation methods for projected area, roundness, and circumference (using Marching Squares)
• New: more prominent Add Pipeline button if there are no pipelines in the document so far
• Fix: preview of analysis operations (e.g. Blob Finder) if the viewer is zoomed out much
• Fix: avoid a potential crash with an empty Compartments analysis operation
• Fix: minor UI glitches in the Compartments' input controls
• Fix: coloring of the output of the Compartments operation
• Fix: potential crash with Configure Output Channels
• Fix: problem on reopening a pipeline with Object Math in 'Merge' mode
• Fix: potential crash with Region Growing's output color configuration
• Fix: avoid error in the Region Growing operation if used with a large number of input objects
• Fix: problem with the labeled image export operation
• Fix: prevent an unwanted change of the current plane on undo of the Input ROI operation
• Fix: incorrect preview for the Enhancement Filter operation 
• Fix: prevent problems in the Intensity Profile while modifying channels with a script

Machine Learning:

• Fix: delay on rename a ML training (objects)
• Fix: selection behavior in the ML Trainer (Objects)
• Improved: error indicators in the ML Trainer (Objects) 
• Improved: don't collapse classes on switching image set in the ML Trainer (Objects)  
• Improved: ML Trainer (Image) together with object styles
• Improved: ML Trainer (Objects) preview together with object styles
• Fix: unexpected stuck with the object classification run
• Fix: avoid problems with parallel running preview operations in the ML Trainer (Objects)
• Improved: tag handling on re-run of the object classification (from trainer panel)
• Fix: more robust test for ML hardware acceleration capabilities
• Fix: no results if running a ML Segmenter operation using 3D features  

Objects:

• Improved: better rounding when calculating feature range for object coloring
• Fix: chart updates object colors after style changes immediately 
• Fix: problem if there are more than 999 tags
• Fix: more reliable tag selection in the remove tags dialog 

4D Viewer & Storyboard:

• Fix: more reliably finish the preview mode after using the storyboard 
• Improved: clipping planes in storyboard movie exports
• Fix: legend more reliable
• Fix: track children visibility in combination with master details view
• Fix: potential missing surfaces in storyboard exports
• Fix: under rare conditions objects were not clipped in storyboard movie exports

Import & Data Handling:

• Fix: broken Luxendo Big Data Viewer import
• Fix: better distinction of mosaic and channel-stitching in LaVision OME tiffs
• New: added GZIp preference switch to the Volume Fusion dialog (for new file results)
• Improved: order of tiles imported from a CZI
• Improved: wording for edit support for tiles in manual mode of the Tile Sorter
• Improved: avoid endless metadata entries in the command history if importing tiled data
• Fix: prevent a potential crash in the Save Image to VisionHub dialog

General:

• Improved: allow object classification with launcher configurations
• Improved: a bunch of minor fixes in the help content
• Fix: wrong GPU RAM display in the System Check tool
• Additional minor fixes