arivis Vision4D 3.6 Release Notes February 2022
Summary of changes in arivis Vision4D 3.6, released 11 February 2022
New & Noteworthy:
- New: Deep Learning inference operations in the Analysis Pipeline
- New: better support for Wellplate data with the new Well Picker panel
- Improved: Blob Finder analysis operation
Deep Learning:
- New: Deep Learning inference with GPU support (CUDA)
- New: use trained deep learning models to do image reconstruction in the Analysis Pipeline
- New: use trained deep learning models in the Analysis Pipeline to segment images
- New: support for the Open Neural Network Exchange (ONNX) format
- New: support for CZMODEL / CZANN Model Specification providing interoperability with ZEN and APEER
Analysis & Objects:
- New: Blob Finder normalization can now be parameterized in more detail
- New: exponential mapping for Blob Finder normalization
- New: allow manual range for Blob Finder normalization
- Improved: Blob Finder progress better reflects normalization step
- Fix: possible problems with Blob Finder preview
- Fix: prevent wrong execution in Blob Finder which leads to strange patterns
- Fix: incorrect progress after stop of the Blob Finder operation
- Fix: surprisingly incorrect results after Split Object
- Fix: save/restore modified object name pattern
Import & Data Handling:
- New: export image data to OME Tiff
- New: Well Picker panel allows to switch between images sets of a well plate easily
- New: Navigator panel shows well plate coordinates for images sets of a well plate
- New: support for ZSTD compression in CZI import
- Fix: rare problems with importing SVS files
- Fix: scenes order for CZI imports
General
- New: arivis is a ZEISS company now
- New: ability to select a specific GPU to be used for Machine Learning
- Improved: better explanation when GPU support for Machine Learning is not available
