MIKAIA FAQs

MIKAIA^® has always been designed for users in life science research and development, who want to be able realize heterogenous analyses without having to develop their own custom software.
Main user groups are academia, CROs, pharma, biotech and biobanks.

MIKAIA^® contains new AIs cleared for diagnostic use. We are happy to point you to a range of companys who focus on this:
– Digital Pathology AI Companies — List of Commercial AIs (CE-IVDR, FDA, RUO)
– The Digital Pathologist’s Cockpit: Image Management Systems (IMS) – Free Download of List of IMSes
MIKAIA^® has always been designed for users in life science research and development, who want to be able conduct heterogenous analyses without having to develop their own custom software.
Main user groups are academia, CROs, pharma, biotech, and biobanks.

MIKAIA^® is a software that you can download and use by yourself in your lab or company.
If you prefer image analysis as a service, we do that too.
See http://www.iis.fraunhofer.de/pharma-services

No, you can purchase it from one of the MIKAIA^® distributors, who license MIKAIA^® from Fraunhofer. We are happy to direct you to the most appropriate partner, based on your territory.

[‘mi:ka:ja:] or “MEE” “KAH” “YAH”
For Germans: Miekaja.

Originally, MIKAIA^® was called MICAIA. The spelling changed beginning of 2024.
MICAIA stands for Microscopy AI Analysis.
The new spelling indicates the German origin: the German spelling is “Mikroskop”.

Until beginning of 2024, MIKAIA^® was spelled MICAIA. We had to change the spelling to settle a dispute with the holder of a somewhat similar trademark.

3DHistech MRXS, Leica SVS, Hamamatsu NDPI/NDPIS, Zeiss CZI, Olympus VSI, DICOM, OME-TIFF (e.g Lunaphore), Akoya QPTIFF, BigTiff, Multifile TIFF (e.g. MACSima), JPG, PNG, TIF
See WSI and Annotation File Format I/O Support in MIKAIA for more details.

No, MIKAIA^® is a native lightweight Windows software and does not even use a central database. You can easily install MIKAIA^® on your computer or laptop, or use the portable version by simply unzipping it. Once installed (or unzipped), you can open slides by dragging and dropping them into the application or by importing one or multiple folders as workspaces. These can be local or networkf folders. 

No, MIKAIA^® uses a decentral scheme: It stores annotations in an *.ano file next to the whole-slide-image file. An *.ano file is actually a zip file and it contains an XML file with the annotations.




This is handy, because it enables users to move the slide (incl. *.ano file) into another folder and the connection between slide and annotation will not get broken. Users can also keep multiple versions of annotations, simply by creating copies of the *.ano file, or create backups of annotation files, for instance after they have created manual annotations.

If you want to store your analysis results separately from the scans, you can easily do that: Just create a shortcut to your original image file and then load the shortcut. MIKAIA will then put the *.ano file next to the shortcut.
When opening the shortcut, MIKAIA^® will open the annotations located next to the shortcut instead of the annotation file located next to the original slide. This way, multiple annotation sets can be maintained, without having to duplicate the (large) WSIs. MIKAIA^® has built-in dialogs that can be used to batch-create shortcuts or retarget existing shortcuts (in case the original WSIs have been relocated).

You can create a separate folder and put shortcuts to your original scans into it. Then load the folder with the shortcuts into MIKAIA. When MIKAIA^® loads an image via a shortcut, it will store annotations next to the shortcut.

Yes, MIKAIA^® includes a tool that lets you redirect to existing shortcuts to a new folder. It is available via main menu | Annotations | Retarget shortcuts in folder. You simply select 1) the folder with the zombie shortcuts and 2) the new folder where the scans have been moved.

Yes, when you’ve annotated or analyzed a set of image scans, the original image files will each have a sibling file with an *.ano extension. Copy all these *.ano files into a new folder. Then use the built-in MIKAIA^® tool at main menu | Annotations | Add slide shortcuts next to *.ano files. You select the folder with the *.ano file copies and the original folder. MIKAIA will then automatically identify matches based on the file names and generate the shortcuts. You can then import the new folder containing the shortcuts + *.ano copies into MIKAIA^®.

MIKAIA^® supports opening annotations in GeoJson format (used e.g. by QuPath), XML format (used by Leica Aperio), or 3DHistech *.dat format. You could also use the REST API to add annotations via a plugin, e.g. writtin in Python.

MIKAIA^® can export annotations to GeoJson format (used by QuPath), XML format (used by Leica Aperio), or CSV format.
Using the Annotation Image Export App or Tile Export App, you can also export images of annotations or export tiles along with masks (e.g., for AI training outside of MIKAIA^®).
When you export a slide (e.g., in a different format, a crop, or at a lower resolution), you can also select to burn in the annotations.

Yes. Even though MIKAIA^® is not a web-based software, it still can be used collaboratively: The free MIKAIA^® lite edition can be used to create annotations manually, but it can also view annotations generated by an analysis conducted with MIKAIA^® studio. In other words, user A can run an analysis with MIKAIA^® studio on computer A, and later users B-Z can view the results on their own computers using the free MIKAIA^® lite edition.

Yes, the Tissue Detection App works out of the box.
The AI Author can be used to train a tissue classifier, e.g., for masking tumor, necrosis, mucus, or any other tissue type.
The H&E Cell AI can be used to segment and classify cells (learn more about H&E analysis using AIs).
The Spatial Clustering App, Cell-Cell-Connection App, Cellular Neighborhood App, or Proximity Analysis App can be used to analyze spatial layout of detected cells.
The Tile Export App or Annotation Image Export App can be used to export small images.

Yes, the IHC Cell Detection App supports segmenting both positive and negative cells. It can also grade cells based on staining intensity. It can detect nuclear, cytoplasmic, or membranous markers and is very flexibly configurable to support a broad variety of different markers.
The Mask-by-Color App can be used to mask regions based on pixel colors. 

Yes, analysis of IF scans is a popular and major use case in MIKAIA^® — from low-plex scans with just one or two markers up to high-plex spatial proteomics scans.
The FL Cell Analysis App can be used to segment and phenotype cells based on marker expression.  
The Mask by Color App can be used to create a mask by thresholding an individual color channel or a combination of multiple color channels.
The Spatial Clustering App, Cell-Cell-Connection App, Cellular Neighborhood App, or Proximity Analysis App can be used to analyze spatial layout of detected cells.
The Tile Export App or Annotation Image Export App can be used to export small images.
Supported formats are Zeiss CZI, Olympus VSI, Hamamatsu NDPIS, QPTIFF (used by Akoya Biosciences), and OME-TIFF (used, e.g., by Lunaphore, Miltenyi MACSima, bioformats, and many other platforms; see list of file formats). You can also open flat images (e.g., JPG, PNG, or TIF) and select that each of the RGB color channels shall in fact be interpreted as a fluorescence channel.

Yes, it can either be configured in different ways:
● The AI can segment nuclei (e.g., “DAPI”, “Hoechst”, or “DNA” channel) and then dilate them by a fixed radius.
● The AI can be fed both a nuclear and a membrane marker. It will then segment the membrane.
● The AI can be fed only a membrane or cytoplasmic marker. It will then segment the outer cell contours.

It does not include apps specifically tailored to these domains. But the FL Cell Analysis App, for instance, can still be used to detect and phenotype cells in suspension.

No, MIKAIA^® supports z-levels (focus planes) and can analyze individual z-levels, but is a 2D analysis program.

Yes, the Tissue Detection App includes TMA-dearraying. Most apps, including the IHC Cell Detection or FL Cell Analysis Apps, will then group detected cells by core, so that per-core statistics are automatically generated.

Yes.

We have not built in a fixed limit. Computer power and memory availability pose limits. We have had users analyze up to 50plexes, but more may be possible. We are happy to assist and provide voucher codes to users who want to experiment with even higher plexity.

MIKAIA^® can export any brightfield scan that it supports to either SVS, BigTiff, ZIF, or DeepZoom. Learn more about WSI and annotation file formats here.
We are working on exporting to OME-TIFF as well.

No, not at this point. We plan to add this feature soon.

No, not at this point. MIKAIA^® must have direct access to the image scan files.

Yes, you can multi-select slides in the MIKAIA^® workspace. As soon as you select more than one slide, the analyze-“Slide” button turns into the analyze-“Batch” button. Once you select more than one slide, the “Analyze Slide” button changes to “Analyze Batch.” You can also create a *.micbat file (refer to the user manual) that lists the apps to be used, the app configuration file, and the selected slides. To start a batch process, simply open this *.micbat text file. All results, including a combined CSV file with results from all slides, will be saved in an export folder.

MIKAIA^® runs on a Windows laptop or PC. The AI based analysis apps benefit strongly from the presence of a CUDA-enabled GPU (CUDA 8 or newer — see below: “What GPU do I need for fast AI interference?”). They do fallback to CPU when no compatible GPU is available, but the apps that use quite large AI models (H&E Cell AI, HE Crypt Segmentation) are really slow in CPU-mode.
The IHC Cell Detection App, FL Cell Segmentation App, or AI Author are fairly usable in CPU mode, but again, they are faster when a CUDA-enabled GPU is available. 
Due to the large size of whole-slide-images, it is recommended to have 16 GB RAM or more.

MIKAIA^® only supports CUDA-enabled NVIDIA GPUs. GPUs by other vendors are not supported as AI co-processors. MIKAIA^® does not require a professional series GPU, but is also compatible with consumer-level gaming GPUs (NVIDIA GeForce).

No, sorry. Only for Windows.

No, sorry. Only for Windows.

No, it can be a Linux or Mac. You just need to be able to install the WIBU CodeMeter runtime (the copy protection framework used by MIKAIA^®), which is available for download at the WIBU website. 

We offer both an installer and a portable version. The installer is an *.exe file and it requires admin rights for installation. The portable version is a ZIP file that must be manually extracted and then MIKAIA.exe can be started from there. This process does not require admin rights.

Both 2K and 4K displays are supported. MIKAIA^® has also good touch-support. For creating accurate annotations, we recommend using a tablet with a stylus pen (we use WACOM tablets). At exhibitions, we sometimes show MIKAIA^® on a touch-enabled digital whiteboard (we use a Samsung Flip 2).

Yes, MIKAIA^® has a REST API. You can easily integrate your own scripts in any scripting language. We provide Python examples that show how to integrate your own AI as well as a Python API that simplifies calling the MIKAIA^® REST API. A plugin can even run in a Docker or on another computer (even Mac or Linux).

Your plugin can use any framework as using the AI model is entirely handled by the plugin and not by MIKAIA^®. This gives the plugin complete freedom in terms of pre- or post-processing the AI input and output. 

MIKAIA^® lite is free. Most image analysis capabilities are locked in MIKAIA^® lite, though. They are only available in MIKAIA^® studio. Many, but not all apps are included in MIKAIA^® studio – some require purchasing an optional add-on bundle. Browse the editions here. 

Yes, get in touch (mikaia@iis.fraunhofer.de) and we are happy to provide voucher codes that let you temporarily use the locked features.

MIKAIA^® has a very fair price. We are happy to put you in touch with one of our distributors, who can give information on pricing or issue a quotation.
The default is a perpetual node-locked license for a one-time up-front free that includes one year of updates & support. We also offer floating or multi-user network licenses. 

By default, we offer perpetual licenses, i.e., once activated, features are available forever, without any follow-up costs.
A license can be node-locked, which means it is locked to a particular PC, or alternatively, we offer single-user or multi-user floating licenses. Here, the activation file is imported on a server in your network. MIKAIA^® can then be installed on any computer in your network or on a (VPN-connected) home office laptop. At startup, MIKAIA^® will automatically detect the license server and, if a user-slot of the single- or multi-user license is currently free, it will start up as MIKAIA^® studio, else as MIKAIA^® lite. The licensing modes are also illustrated in the MIKAIA^® product brochure.

No, MIKAIA^® is designed for life science researchers. Most apps are quite generically usable. The AI Author is a good example: Users can train their own AIs directly in MIKAIA^®.
MIKAIA^®, incl. all apps, is for research use only (RUO).

MIKAIA^® contains a funtionality that helps with documentation and ensures repeatability. When batch-analysing a cohort, the results folder will contain shortcuts to the analyzed slides, the app settings file that contains all parameters that were used in the particular anylsis, a *.micbat file that lists the entire batch job (and can later be executed again), and the CSV file with all quantitative results. This additionally states the software version, computer and user name, and date.

MIKAIA^® works offline. The only communication is at startup, when by default, MIKAIA^® checks with a Fraunhofer server if a new version is available and, if that is the case, then shows a notification dialog. This check can optionally be disabled and has no further impact of the functionality of MIKAIA^®.