For each application, we use precisely trained AI models that enable automated preprocessing of metallographic micrographs and consistent, standard-compliant quantification — robust against variations in preparation, image source, and user behavior.

Our system currently offers a powerful module for determining the grain size of single-phase materials. It was originally trained on metallic structures, but can be reliably generalized across a wide range of different material classes.

For more complex applications — such as those involving special structural features or insufficient standard performance — our models can be specifically adapted with minimal training effort. Just a few additional image examples are sufficient to develop robust, tailor-made evaluation models for your specific requirements.

In addition, we offer a specialized module for comprehensive phase and grain boundary analysis of two-phase steels with a ferritic matrix and a carbon-rich second phase.

Our AI-supported evaluation modules relieve qualified specialists of time-consuming routine tasks — a decisive advantage in view of the increasing shortage of skilled workers. This allows your experts to focus their capacities on more demanding tasks and decision-relevant assessments.

At the same time, processing times are significantly reduced while the quality of the analysis increases: AI enables detailed and consistent recording of complex structural features — objectively, in compliance with standards, and in a reproducible manner.

Another key benefit is independence from the individual user: variations due to individual experience or interpretation are eliminated, which significantly increases the comparability and reliability of the results — whether in quality assurance, development, or series approval.

After uploading your microstructure images via the web app, select the appropriate analysis module and the calibration factor that matches the microscopy configuration used. These can be stored once to speed up the selection process.

The evaluation is automated — the result is available shortly thereafter in the form of a standardized results report. In addition to classic microstructure parameters, this report also includes statistical distributions of relevant characteristics. The system also generates a visual representation of the individual quantification steps so that the analysis can be reviewed by expert users and checked for plausibility.

Our web app is designed to be ready for immediate use without local installation — regardless of location or existing infrastructure. Evaluation is currently performed by uploading previously captured micrographs (e.g., from SEM, light, or laser microscopy), which can be stored in the usual manner and then processed directly via the platform.

For even closer process integration, we are currently developing an API interface that will enable direct connection to existing systems and automated data transfer in the future.

No — our web app was specifically developed for practical use without any programming knowledge. The user interface is intuitive and guides you step by step through the analysis process — from image upload to results report.

For an optimal start, we are happy to offer accompanying training courses. In these sessions, we not only teach you how to use the web app, but also demonstrate the functional possibilities of our evaluation modules.

The methodology and structure of our evaluation modules are based on the relevant standards for metallographic microstructure analysis, in particular DIN EN ISO and ASTM standards. This applies, for example, to parameters such as grain size, phase distribution, and morphology.

However, it is important to emphasize that current standards do not contain specific regulations for AI-supported processes. Formal “standard compliance” in the narrower sense is therefore not currently possible.

Our AI-supported microstructure analysis provides standardized and objective evaluations — regardless of the operator, location, or image quality. This significantly reduces deviations and room for interpretation, which often occur with manual procedures.

The consistent database minimizes the coordination effort with your customers, reduces the likelihood of complaints, and builds trust through traceable, reproducible test reports.

Yes — our platform is fully scalable and can be flexibly adapted to a wide range of requirements. The web app allows multiple micrographs to be uploaded and automatically evaluated in a single step — ideal for series testing or larger sample sizes.

The browser-based architecture enables location-independent access without local installation or hardware binding.

The question of reliability can never be answered with absolute certainty in microstructural evaluations — neither for conventional methods nor for AI-based approaches. This is because structural analysis is also subject to system-related uncertainties, such as image quality, preparation, or sample size.

Our AI models have been trained with the utmost care using methodologically sound and diverse data sets — with the aim of achieving not only high precision but, above all, robust generalizability.

Data security is our top priority — especially in the sensitive field of materials analysis. All data transfers and storage are encrypted and subject to strict access controls.

Your data is stored exclusively on local servers in Germany — fully GDPR-compliant and in accordance with current security standards. By default, only the respective user or customer has access to their own data.

No — quite the opposite. Our AI supports you in areas where human working time was previously tied up in repetitive, manual routine tasks. Automated evaluation lightens the load, but does not replace you: your professional expertise remains indispensable — especially when it comes to interpreting, evaluating, and approving the results.

AI is a tool, not a substitute for experience. It serves to standardize, accelerate, and safeguard — not to replace expertise.

First, it is important to emphasize that AI is not a panacea. Our models have been trained on the basis of diverse structural patterns — nevertheless, in certain cases, the results may not meet expectations.

If you have individual images or evaluations that the system does not work with satisfactorily, please contact us. We take such cases very seriously, investigate the cause together with you, and offer to further train the existing model with a few representative examples.

Yes — depending on the analysis module used, so-called expert parameters are available that allow you to intervene specifically in the evaluation process. These options were designed specifically for demanding or atypical microstructures that require finely tuned control of the AI evaluation.

By adjusting these parameters, you can influence segmentation sharpness, threshold values, or detection limits, for example, in order to optimize quantification in challenging cases.

Yes — we offer the option of developing custom-trained AI models that are precisely tailored to your materials, questions, and analysis goals. While our generic standard models are designed to offer a good balance between generalizability and accuracy across many material classes and microstructures for various contrast and imaging techniques, customized models may be superior in specific use cases.

An individually trained model behaves similarly to a subject matter expert who specializes in a specific microstructure.

Our customized AI solutions are developed in close collaboration with you: Together, we clarify the task at hand, the constraints, and the target parameters. Based on reference samples and using correlative microscopy (e.g., SEM, EBSD), we create a reliable, objective ground truth — even for microstructural details beyond the scope of light microscopy.

These findings form the basis for a trainable data set that we use to develop an AI model that subsequently delivers reproducible results based on light microscopic images.