ZEISS Knowledge Base

Toolkit/System	Features/Improvements
Base Software	General: Expanded the AI Model Store to allow filtering of models by name. Included Docker Desktop in the ZEISS Microscopy Installer. Improved Smart Setup by adding new dyes from the FPbase.org database. Updated the arivis Pro bridge to allow sending czi images to arivis Pro without a conversion. Removed the arivis Cloud (on-site) functionality. Direct Processing: Added Lightfield 4D Processing.
Airyscan	Expanded Airyscan Joint Deconvolution for support of Airyscan Multiplex.
Bio Applications	Changed the designation of the toolkit formerly known as Bio Apps toolkit to Bio Applications.
Celldiscoverer 7	Added support of additional multi-well plate formats for reliable sample carrier calibration.
Developer Toolkit	Migrated to Iron Python 3.4 which introduces syntax changes that may render existing macros incompatible. To mitigate this, a conversion tool is provided to update scripts for compatibility with ZEN 3.12 and higher. You can convert macros in .py format, e.g. using ZeissPython directly with the 2to3 tool. Scripts in .czmac format can be converted with the ZeissMacroConverter.py and ZeissMacroBatchConverter.py batch conversion script (available at the ZEISS Github page). For more detailed information, see Breaking Changes for Developer Toolkit and Smart Acquisition Toolkit.
FLIM Application	Enabled FLIM acquisition for NLO tracks with PicoQuant detectors.
Lightfield 4D	Introduced Lightfield 4D to image an entire volume at an exact point in time without any time delay. Instead of capturing single 2D images at different time points, a micro lens array positioned between objective and camera generates 37 individual images and collects all the 3D information at the same instant.
Microscopy Copilot	Added Microscopy Copilot, an AI assistant designed as a conversation tool helping LSM users to set up imaging experiments on LSM systems.
Slidescan	Created a new version of the macro for custom sample detection. The version number is part of the name and indicated in the macro code as a comment. Introduced a new barcode engine. Added support for Axicam 820 mono with 0.8x adapter for Spatial Biology. Enabled Axicam HDR mode with 16 bit output.
Smart Acquisition Toolkit	Guided Acquisition Implemented the possibility to loop and repeat the entire Guided Acquisition workflow and/or detailed scans allowing dynamic observation of rare events over time. Experiment Feedback Migrated to Iron Python 3.4 which introduces syntax changes that may render existing macros incompatible. If necessary, modify your existing Experiment Feedback scripts to be compatible with Iron Python 3.4. For more detailed information, see Breaking Changes for Developer Toolkit and Smart Acquisition Toolkit.
Spatial Biology	Introduced a new toolkit initially only available with the special Spatial Biology configuration of Axioscan 7. Includes Low Contrast AI Tissue Detection, Co-Registration, and FINE functionality. Low Contrast AI Tissue Detection adds a new tissue detection method to the Axioscan 7 workflow which is capable of detecting faint contrast of fluorescent tissue in fast brightfield (flash) prescans, combining speed and robustness. Co-Registration adds a new Tissue Image Alignment image processing function for combining multi-channel images of different staining and imaging cycles. Fluorescence Intensity Normalization and Evaluation (FINE) ensures the same fluorescence intensities across multiple scanners once set up in factory.
ZEN API	Introduced ZEN API as a open interface designed to control or systems from external applications.

Breaking Changes for Developer Toolkit and Smart Acquisition Toolkit

For the Developer Toolkit (OAD) and Smart Acquisition Toolkit (Experiment Feedback) there are possible breaking changes to scripts due to upgrade from IronPython 2.7 to IronPython 3.4. This upgrade introduces syntax changes that may render existing macros (.czmac or .py) incompatible with the new version.

For Experiment Feedback scripts you have to manually adapt the scripts to be compatible with IronPython 3 in case they are incompatible.

For other existing macros that will no longer work due to syntax differences between IronPython 2 and IronPython 3, a conversion tool is provided to update existing scripts for compatibility with the latest ZEN versions. To convert your macros you can either use the ZeissPython distribution which you can find in the ZEISS Microscopy Installer (ZMI), or you can use your own python distribution. You need the 2to3 tool, which reads Python 2.x source code and applies a series of fixers to transform it into valid Python 3.x code.

If you use the ZeissPython distribution available with the ZEISS Microscopy Installer, the 2to3 tool is located at located at C:\Program Files\Carl Zeiss\ZeissPython\Py23279.1\env. If you use your own Python distribution to run the conversion tools, you have to ensure that the 2to3 tool is available in your Python distribution. If necessary, it can be added using: pip install 2to3

If you are operating older and newer ZEN versions, you may need to use two different versions of your OAD scripts, one version that is compatible with IronPython 2 and another version of your script that is compatible with IronPython 3.

Converting OAD Macros in *.czmac Format

The python 2to3 tool cannot be applied directly to ZEN macro files in *.czmac format since they are XML files that contain Python code in one of its XML elements. For these macros we have provided a script on the ZEISS Github page which automatically converts an input macro (*.czmac) file from python2 to python3 syntax.

The 2to3 tool is available on your PC.

Go to the ZEISS Github page and download ZeissMacroConverter.py. To batch-convert all macro files in a specified directory, additionally download ZeissMacroBatchConverter.py. If you download the batch file, store it at the same location as the other file.
The files for conversion are available.
Use the command line and call the python script from command line with the parameters -i (with the path to the file that needs to be converted) and -o (with the path to where the converted file needs to be stored, which can be the same as input file path). Example for converting a script called analysis time.czmac: D:\>python "D:\temp\macro\ZeissMacroConverter.py" "D:\temp\macro\analysis time.czmac" -o "D:\temp\macro\analysis time_converted.czmac"
The respective macro is converted.
To batch-convert macros with the ZeissMacroBatchConverter.py, you need to specify the path to the root folder containing the OAD macro files, for example: D:\>python "D:\temp\macro\ZeissMacroBatchConverter.py" "C:\OAD_Macros"
All macros in the folder are converted.

Special Notes

There are some special cases which are not automatically converted. One of them is concatenating "str" and "bytes", in python2 this works, but in python3 you need to explicitly convert (encode/decode) one of them so that they are both "str" or "bytes"!
Problem with trying to access a member called "None": In Python 3 you are not allowed to use access members called None. The example conversion macros linked above take care of this problem by using regex to replace the problematic parts as suggested here: https://github.com/IronLanguages/ironpython3/blob/v3.4.0/Documentation/upgrading-from-ipy2.md
This is an example of the old and new syntax:

#IronPython2 syntax:
set = Zen.Processing.Filter.Settings.GaussSetting()
set.Dimension3rd.None
#IronPython3 syntax:
set = Zen.Processing.Filter.Settings.GaussSetting()
set.Dimension3rd["None"]

Compatibility of Windows Versions to Machine Learning and AI Features

ZEN allows you to use certain machine learning and AI functionality, e.g. semantic segmentation and denoising (via the respective Intellesis functionality) as well as the use of AI models for instance segmentation.

The following table illustrates CPU and GPU usage for ZEN Toolkit AI, ZEN Toolkit 2D, and ZEN Toolkit Bio Applications:

	Windows 10 version 1607 (build 14393), e.g. Windows 10 Enterprise LTSB 2016	Windows 10 version 1809 (build 17763), e.g. Windows 10 Enterprise LTSC 2019	Windows 10 version 21H2 (build 19044), e.g. Windows 10 Enterprise LTSC 2021	Windows 11, version 24H2 (build 26100), e.g. Windows 11 Enterprise LTSC 2024
ZEN Toolkit AI	CPU fallback for semantic segmentation and denoising.	Additional GPU support for semantic segmentation and denoising.
ZEN Toolkit 2D ZEN Toolkit Bio Applications	CPU fallback for semantic segmentation.	Additional GPU support for semantic segmentation and instance segmentation using CPU.	Additional GPU support for instance segmentation.

Windows 10 version 1607 (build 14393), e.g. Windows 10 Enterprise LTSB 2016

Windows 10 version 1809 (build 17763), e.g. Windows 10 Enterprise LTSC 2019

Windows 10 version 21H2 (build 19044), e.g. Windows 10 Enterprise LTSC 2021

Windows 11, version 24H2 (build 26100), e.g. Windows 11 Enterprise LTSC 2024

ZEN Toolkit AI

CPU fallback for semantic segmentation and denoising.

Additional GPU support for semantic segmentation and denoising.

ZEN Toolkit 2D

ZEN Toolkit Bio Applications

CPU fallback for semantic segmentation.

Additional GPU support for semantic segmentation and instance segmentation using CPU.

Additional GPU support for instance segmentation.

When using instance segmentation models trained on arivis Cloud for image segmentation in ZEN Toolkit 2D or ZEN Toolkit Bio Applications, an installation of Docker is necessary. Depending on your Windows and Docker version, the functionality can be limited to only run on CPU instead of GPU. The following tables illustrate which Docker and operating system version you need if you want to use the functionality also with GPU, and if you need to use Hyper-V or WSL.

To install Docker Desktop, you can use the ZEISS Microscopy Installer, with the exception of older Windows versions, for which you need to install an older version of Docker Desktop directly from the Docker website, see table below.

	Windows 10 version 1607 (build 14393), e.g. Windows 10 Enterprise LTSB 2016	Windows 10 version 1809 (build 17763), e.g. Windows 10 Enterprise LTSC 2019	Windows 10 version 21H2 (build 19044), e.g. Windows 10 Enterprise LTSC 2021
Docker version for instance segmentation	not supported	Version ≤ 4.24.2 For version 4.24.1, use https://desktop.docker.com/win/main/amd64/123237/Docker%20Desktop%20Installer.exe	Current version available with the ZEISS Microscopy Installer
WSL/Hyper-V	not supported	Hyper-V	WSL 2
ZEN Toolkit 2D	No instance segmentation possible.	Instance segmentation using CPU.	Additional GPU support for instance segmentation.
ZEN Toolkit Bio Applications	No instance segmentation possible.	Instance segmentation using CPU.	Additional GPU support for instance segmentation.

Compatible Cameras

The table below displays which cameras are compatible to the software.

Compatible ZEISS Axiocams

Camera	Connection
Axiocam 105 color	USB 3.0
Axiocam 202 mono	USB 3.0
Axiocam 203 mono	USB 3.0
Axiocam 208 color	USB 3.0
Axiocam 212 color	USB 3.0
Axiocam 305 color / mono	USB 3.0
Axiocam 503 color / mono	USB 3.0
Axiocam 506 color / mono	USB 3.0
Axiocam 512 color / mono	USB 3.0
Axiocam 702 mono	USB 3.0
Axiocam 705 color / mono	USB 3.0
Axiocam 712 color / mono	USB 3.0
Axiocam 807 color / mono	2x USB 3.0
Axiocam 820 color / mono	2x USB 3.0

Other, compatible cameras

Camera	Notice
PrimoStar HD	USB only, not via network connection
Primovert HDcam	USB only, not via network connection
Hamamatsu: Orca Flash 4.0 V3 CL/USB3.0, Orca Flash LT USB3.0, Orca Fusion, Orca Fusion BT
Roper Photometrics: Prime 95 B, 95B 22mm, Prime BSI, Prime BSI Express
PCO: edge 5.5 CL, edge 4.2 CHLS USB3.0, edge 4.2 bi, edge 4.2 CL, edge 4.2 LT

Carl-Zeiss-Strasse 22

73447 Oberkochen

Germany

ZEN 3.12

New Features & Improvements