At Evident's European headquarters in Hamburg, Markus Fabich, Senior Manager Strategic Marketing, and Enrico Poege, Global Marketing Communications Lead, present the AI functions of the Preciv software and the new DSX2000 digital microscope. Central to both is the extensive support for operators, whether through automation or AI assistance functions. In addition, a product innovation was presented which will only be introduced to the public in May of this year.

inspect: For those readers who are not familiar with the Preciv software: What are its key features?

Markus Fabich: The Preciv software is a system solution for microscopic analysis and measurement technology in research, development, and quality control in industry. The field of application ranges from micrography to critical dimensions in circuit boards and failure analysis, from series analysis to individual testing.

inspect: Is the software only compatible with Evident microscopes?

Fabich: The software is specifically designed for Evident microscopes, in particular for the DSX digital microscope series. But it can also be combined with other microscopes, provided that a compatible camera is used. However, in this case, the microscope must be a manual or motorized device with limited image calibration functions.

inspect: What are the main applications in which your customers use Preciv?

Fabich: These are, of course, areas where we are definitely one of the market leaders, including classical metallography. We use specific and proven methods there, for example, to analyze porosity. But applications that involve geometric tolerances and dimensions are also included, such as measuring the cross-section of conductor paths on printed circuit boards. And then there are, for example, extensive analyses for powders; such as size class determination for additive manufacturing.

inspect: Preciv has now also been given AI functions. What can users expect from these?

Fabich: With our AI model, we are focusing on our customers' applications. And, as I said, these are pattern recognition and geometry measurement. The focus is on defining “good” and “bad” samples. This is called “ground truth”. Labeling images is a useful way of creating this. And that then leads to the actual pattern recognition, i.e., a circle is recognized as a circle relevant for the geometry measurement. And in the second step, this circle is assigned to its corresponding size class.

inspect: Who does the training? inspect or the user?

Fabich: The user. The simple reason for this is that many of our customers generate highly confidential data that we have no intention of accessing. Of course, we could offer pre-training as a service. But the autonomy of the user is our top priority. They can determine what happens to their data and be sure that only their data is used in their solution. Logically, this also implies a requirement for the usability of the software: the training must be designed in such a way that the user can easily carry it out themselves.

inspect: Let's assume that the AI model is trained. How does the user benefit from this?

Fabich: The AI can then be activated at various points in our software, for example in the live image. Then, for example, it measures the diameter of a circle and marks it green if it is the right size, or red if not.

inspect: With your Easy Mode and the convenience features, some of which are also AI-supported, are you addressing the shortage of skilled workers?

Fabich: We have heard from customers in the field of quality assurance that they are having difficulties finding specialists with sufficient knowledge of microscopy. I believe that the DSX is a device that can be learned relatively quickly without much training – simply because so much is automated. Especially the version that is equipped with the automatic objective revolver. This can also be taken further: an engineer can create the workflow with the relevant process steps. In Easy Mode, the operator then simply works through them. The operator then only sees the previously defined functions.

Enrico Poege: In addition, the influence of the individual operator on the measurement result is minimized in this way. This is important for comparability.

inspect: Preciv has the Live Assistant function? What is its specific benefit?

Fabich: Provided that the user has already trained the AI with his images, he can now count particles or defective drill holes in the live image, for example. The result always refers to the current field of view and changes dynamically. This means that if I move the sample, the number of visible defective drill holes would always be displayed.

inspect: Preciv now has a Microsoft 365 integration. This means that you can export measurement results to word processing and spreadsheet programs. Can you explain this a little further?

Fabich: As you know, it's not enough to take a good picture and take a measurement. Rather, a report has to be created. And a plugin for Microsoft 365 helps with that. This allows the user to create a template for which placeholders for images and tables can be predefined. As soon as I click on “Create report in Word”, Office opens automatically, and then I can export the images from the gallery and the measurement value tables to their corresponding locations using drag and drop. It is also possible to create scales or diagrams in PowerPoint, so that the user does not have to create them from the data themselves.

inspect: Let's move on to the new DSX2000 digital microscope. What are the key data and key features here?

Poege: With the DSX2000, we now have a fully automated digital microscope. The most important innovation in this respect compared to the DSX1000 is the objective revolver, which can be controlled automatically. This of course makes things easier for the user.

Fabich: What is unusual about this nosepiece is that it can also be used in a tilted position. And if I have done that, it is factually correct that the rotation via software is no longer allowed, because the objective in the tilted position could possibly collide with the carrier plate. But what is so charming about this nosepiece is that I can still move it into position by hand. So the automatic and manual processes are combined. This is something we didn't have before.

inspect: What is the advantage of a digital microscope over an analog one?

Poege: A digital microscope makes work easier for the user through automation – for example, integrated contrasting methods – that can be applied with just one click. Many functionalities, including the wizards mentioned, can be accessed directly from the console. Furthermore, you can switch back and forth between the different modes on the console; for example, from DIC (differential interference contrast) to bright field, depending on your current needs, without having to exchange optical components.

inspect: Let's take a step away from digital microscopes: Where are analog microscopes mainly used today?

Fabich: There are specialized applications, such as infrared microscopy – the sensor in the DSX2000 does not support infrared. But even if only an inexpensive camera is needed and only one light method is used anyway, a component-based microscope can be the optimal choice for standardized workstations. Even for specific applications, such as wafer inspection, a classic system may be the better solution. In this case, the MX63 with its large, stable stand may offer more advantages. The versatility of modular microscope systems, combined with the ability to view samples directly with the naked eye, remains relevant for such special applications.

inspect: The DSX2000 features the so-called Best Image function. What exactly is meant by this and how does it work?

Poege: This function primarily gives the operator certainty when using the contrast methods. He no longer has to know exactly which lighting technique is best in which situation. But the software goes a step further: it automatically checks the available light techniques and the operator can select the image in which the feature to be examined is best seen with a single click. This avoids the situation where it is determined after the fact that a different light technique would have been better. The same applies to contrast, brightness, polarization, and other parameters.

Fabich: I would also like to add that once the optimal setting has been found, it can be saved. This means that the same parameters can be used again for other samples. The best image function thus helps to find the ideal mode for a particular sample.

inspect: What will Evident be presenting to the public next?

Fabich: We are developing the DSX into a comprehensive ecosystem. The first accessory is a macro camera that is equipped with a ¼-inch thread and can be mounted on the side of the microscope or in another suitable location. The cable is long enough for flexible use. In addition, a new operator workflow called “Macro Camera” is being integrated into the software. In failure analysis, for example, the camera makes it possible to pre-focus a sample before placing it under the microscope in order to find the region of interest more quickly. Alternatively, it can also be used as a standalone macro station, in which case it takes on the tasks of a mobile camera. The macro images created are available directly in the DSX ecosystem and can be seamlessly integrated into reports. At the Control trade fair in Stuttgart, the macro camera is our special highlight.

AUTHOR
David Löh
Editor-in-Chief of inspect