The author of RoScan is Adam Chromý from the Cybernetics and Robotics group at CEITEC BUT, who also works at the FEEC Department of Control and Instrumentation. His robotic 3D scanner differs from similar devices in its unusual constitution, and it can also be combined with sensors such as a thermal imaging cameras. After combining the image with a colour 3D model, it can not only detect, for example, an ongoing inflammation, but also locate it. It helped diabetes patients at the St. Anne’s University Hospital Brno, and it can also be used in construction, mechanical engineering, chemistry or electrical engineering.

Doctor, can you describe your device in simple terms?

RoScan is in itself a robotic 3D scanner, so its function is to scan an object from the real world and create a computer model that can be virtually viewed, rotated, etc. The very possibility of transferring an object from the real to the computer world is not so special, because commercial 3D scanners are now quite widespread across various industries. However, what is interesting about RoScan is its unusual constitution, thermal imaging layer, and especially its multipurpose application.

What makes RoScan’s constitution unique?

To put it very simply, 3D scanners are mostly handheld or railed. Handhelds have the advantage of being able to scan even complex shapes, but they are less accurate because the scanning head does not know exactly where it is and has to calculate based on its measurements. On the contrary, those on the rails are very accurate, but their scanning trajectories are limited, so they cannot capture, for example, hollow objects and such. We used a robotic arm that is very flexible and very accurate, so we were able to eliminate both disadvantages of current solutions. Ten years ago, when we started the project, we were one of the first to think of something similar.

So what are its main advantages?

Nowadays, it is common for a 3D scanner to obtain a three-dimensional computer model in colour. We achieve this result by having a laser rangefinder on a robotic arm which we can move. We always know exactly where the arm is located, in which direction the rangefinder is looking and at what distance. Thanks to such data, we can calculate where the measured point is located in space. We repeat this countless times and a 3D model is created. Subsequently, we map an image from a colour camera to its surface, thus obtaining the same result 3D scanners would achieve. However, unlike in their case, we can very easily add any other sensor – such as a thermal imaging camera, hyperspectral camera, UV camera or anything else we need for a scientific experiment, and then display this data directly on the surface of the model. And that is the greatest advantage of this constitution – that universality.

You mentioned the advantage of incorporating a thermal imaging layer. What makes it so interesting for you?

Thermal imaging is a camera that shows you the temperature for each point of its image, which is very useful in various fields. For example, in construction, it lets you know where heat is escaping; in electrical engineering, you can the largest current flows; and in healthcare, it can point out inflammation or badly perfused tissue. However, it has several major disadvantages that limit its usability. The first of them is the problem of localisation, since you do not know from the thermal imaging image exactly where the given problem is. You know you have a warmer (inflamed) spot on your skin, but it's not clear if it is related to a sign, a scratch a little next to it, or something else entirely. However, when we combine such an image with a colour 3D model, we immediately see that it is, for example, the sign. This is very useful, for example, in healthcare, where thermal imaging is used very little compared to its potential.

Few people have experience with specialist examination using a thermal imaging camera. Why do you think we are failing to make better use of its potential in healthcare?

The literature indicates three main limits of medical thermography: in addition to the already mentioned localisation, these are insufficient resolution and poor reproducibility of experiments. Although technology has developed significantly over the last few years, the camera in the cheapest mobile phone still gives you a picture in a higher resolution than the most expensive thermal imager. Because the resolution dictates the detail of the resulting image, the doctor must choose whether to see the patient’s entire leg in the image, but only very roughly, without details, or, conversely, a very detailed image of a tiny section. However, what the doctor needs is a very detailed picture of the whole leg, which just cannot be delivered by a classic thermal imager. However, we can deliver that, because we can combine many detailed images on the surface of a 3D model into one large image without any distortion.

That looks interesting, and what about that reproducibility?

The problem with reproducibility is that it is difficult to place the patient in front of the lens in exactly the same position as in the previous scan, so you cannot compare the changes between images in detail. However, once you have projected all the images on a 3D model of the patient, comparing the two images is no longer a problem.

If I understand it correctly, RoScan is the ideal solution in this regard…

No, the ideal solution is something that does not exist and we can only get closer to it. It is true that RoScan really eliminates all three major problems of medical thermal imaging, thus opening up space for its wider use in medicine. This makes it a little closer to the ideal solution than the current solution.

How did the idea to design a similar device come about? Is it the result of a project?

The idea was born about ten years ago in Professor Žalud’s head as a topic for my dissertation. Since then, it has been developed more or less to this day. The biggest shift in its development occurred within the European project ASTONISH, the output of which were six new diagnostic devices for medicine, including our RoScan. It was a project that expert evaluators described as the most successful project they had ever assessed. In addition, it helped me successfully defend my dissertation.

Who is involved in the development of RoScan?

Given that I have been associated with this project since my studies at FEEC, much of the development will probably be associated with my person. However, almost everyone in the Robotics and Artificial Intelligence group who worked with me on this project is bound to have a share in RoScan. And of course also the members of RG2-2 CEITEC, because without their infrastructure, this research would be an immensely difficult task.

What stage of research are you at?

The device itself is in the prototype phase and we use for our research projects. The very core of the robotic 3D scanner is more or less ready and we are just modifying it for the specific application we want to research. However, there are a number of those applications, so the research itself is definitely not over.

You mentioned that your robotic scanner has a multipurpose application. What are the other areas and purposes of its use?

Wherever we need to measure the temperature without contact and we are not only interested in the temperature of the object as a whole, but especially the temperature distribution in its individual parts. Therefore, it can be used in construction, mechanical engineering, chemistry, electrical engineering, healthcare, environmental care and many other industries.

Is your device already used in practice? If so, where can we see it?

RoScan was last put into practice about a year ago at the St Anne’s University Hospital Brno, where several dozen diabetic patients passed under its scanners. This study showed that the measured data can help even less experienced staff to make highly expert decisions. Since then, however, only Covid-19 has been the topic in our hospitals, and there is no time or energy left for any further experiments. We hope that will change soon. We are currently looking for other partners from clinical practice to whom we can offer help with the issue of objective medical quantification.

Where would you like to see RoScan in the future?

I would like this device to be used anywhere it makes sense. I see the biggest use-case in healthcare, where the application of RoScan could provide doctors with objective information about the patient’s condition. They could then make better and more accurate decisions, which would bring faster and more effective treatment, moreover at significantly lower financial costs.


Source: vut.cz/en