Visual computing from Vienna among the world’s best

In Vienna’s Donaucity, people are working on the latest technology in the field of visual computing – and walking on Mars.

With VRVis, Vienna has an internationally recognised research facility in the field of visual computing. More than 75 researchers at the Ares Tower in Donaucity are working on the technologies of tomorrow, which go far beyond conventional virtual and augmented reality. They create customised technology solutions in close cooperation with partners from science and industry.

VRVis is a COMET competence centre (“Competence Centres for Excellent Technologies”) for visual computing. What does that mean?
Gerd Hesina: Visual computing is an area of computer science that brings together many scientific disciplines that have to do with visual components. This ranges from computer graphics, computer vision, visual data analytics, artificial intelligence, extended reality, image processing and simulation to digital twins. In visual computing, people’s eyes are used to carry information. If you think of the eyes as a broadband connection that we use to take in information, it’s the fastest way into the brain.

One of the areas is image processing, which is about decision- making based on images.
Hesina: We use image processing and image processing algorithms, enriched with artificial intelligence, to create possibilities that were only dreamed of a few years ago. We have been working on this for more than ten years and have been able to implement many projects. We are very strong in this area in the development of AI-supported trustworthy diagnostic solutions for digital radiology. Here we focus on the combination of visualisation and artificial intelligence for better comprehension and for greater reliability of AI results. This combination of visualisation and AI is a real power couple, making satellite imagery more usable for sustainable digital agriculture, taking flood simulations to a whole new level, or driving many innovations in working with point clouds in the fields of infrastructure and construction.

VRVis sees itself as a bridge between research and business. Can you give us an example?
Hesina: I wouldn’t say that. We don’t have Google or Amazon research in Vienna, that’s true. But what we do have in Vienna are special areas in which there is traditionally a lot of research, and one of them is computer graphics, which has grown a lot over the years. This development began in the 1980s and 1990s. VRVis also emerged from this environment in 2000. Together with University of Technology, the Graz University of Technology and the University of Vienna, we form one of the largest research clusters in Europe in the field of visual computing.

VRVis sees itself as a bridge between research and business. Can you give us an example?
Hesina: We ensure a lot of knowledge transfer. If novel methods have been developed in science, that does not mean that they have already been implemented in an application or are ready for use. That means we take methodologies and algorithms and mould them into application-oriented software. An example of this would be the flood simulation that VRVis developed for the Federal Ministry of Agriculture, Regions and Tourism and the National Insurance Association for the whole of Austria. Here we have recalculated the flood risk zones in cooperation with the Vienna University of Technology and show them interactively to make the issue more accessible for people. Based on this, emergency plans are created for heavy rain events and floodings, and scenarios are calculated, which suggest measures ranging from sandbag barriers to mobile protective walls. Our flood simulation is also very well received in Germany, for example in Hamburg, Cologne, Rhineland-Palatinate and in East Frisia – that is knowledge transfer directly from science into an application

Another focus is on augmented and virtual reality. Will we all soon be walking around with VR glasses?
Hesina: Even in the 1990s, the prevailing opinion was that VR technology would make a breakthrough and that in the future people would walk around with VR helmets or VR glasses. That has not happened. Improved hardware has created completely new possibilities that are broadening the range of applications. Today, VR glasses are used for training purposes, for example, for fire protection training. In this case, the VR goggles provide a realistic representation of a fire scene without igniting fires or having to waste expensive extinguishing agents. I can experience the ‘look and feel’ of the training virtually in exactly the same way and thus practice realistically for an emergency.
We also work with retirement homes and eye specialists to make potential visual disorders and their implications in everyday life easier to understand, or with the Austrian Armed Forces in the field of remote sensing, where soldiers virtually explore distant terrain. In the same way, our 3D viewer PRo3D, which is well established in planetary research, allows geologists to walk on Mars as if they were there in real life and to examine the rock layers with precision down to the millimetre

You also deal with reconstruction. What exactly does VRVis reconstruct?
Hesina: We specialise in a wide variety of application areas, because many fields benefit greatly from 3D reconstructions. A common denominator of our projects is that we develop visualisation- based software solutions that massively simplify the work with complex point cloud data. In concrete terms, this works by using laser scanners to translate real objects into a digital 3D model, a so-called digital twin, which makes planning processes and simulations very easy. Incidentally, we also use photogrammetry for this purpose, where we develop our own software to create 3D models made from photographs.

Where are your simulations used, apart from flood scenarios?
Hesina: Simulations have long since arrived in everyday life. It can be summarised under the term Smart City, for example when we simulate traffic flows in cities. Different sensors collect data, build a mathematical model and create an overall picture. This could be an evacuation scenario at a football match in a stadium, to name just one example.

How far has the development of artificial intelligence come?
Hesina: When you talk about AI, you should ask what’s under the bonnet. There is still a mathematical model behind it that gets its data from different sources. I can’t invent information, and neither can artificial intelligence. It may be that in ChatGPT, for example, apparently invented information occurs because it was trained that way in the model. But the model itself did not invent it. We are not talking about real AI here. It will always require a human to control the information.

So, there is no reason to worry about computers taking over the world?
Hesina: No, of course not. It always depends on how systems are linked and used. But if you become totally dependent on one system, then you will probably be dominated by it at some point.

What role does sustainability play at VRVis?
Hesina: This is an extremely important topic for us. We have worked very intensively on the UNO Sustainable Development Goals, of which there are 17. We are proud that we are already covering ten of the 17 development goals in implementation. These include topics such as accelerating radiological diagnostics, visual analytics for the energy industry, human-centered analysis tools for renewable energy, trustworthy AI in biomedicine, water-sensitive urban planning for sponge cities and blue-green infrastructure, autonomous, AI-based construction site documentation, 3D visualization tools for space research and geology, and much more. Of course, we are not producing photovoltaic systems or heat pumps, but we use simulations to calculate what the energy balance of a building looks like, for example, and help to build planning offices sustainably and cost-effectively. And we are researching smart climate change adaptation strategies, where visual analytics, among other things, plays a major role. This involves optimizing power plants, faster maintenance of turbines or the use of 3D printing in rail transport to produce spare parts more sustainably.

A number of spin-off companies have already emerged from VRVis’ activities.
Hesina: Yes, we are very proud that start-ups have emerged from several of our research projects. These are of course, like patents very important for the continuation and establishment of research results in practical applications. Our most recent spin-off is Visplore, a visual analytics tool developed at VRVis that makes the analysis of large, heterogeneous sensor data from the industrial and the energy sector intuitive and more efficient.

Why is there a shortage of IT specialists in Austria? Do you feel it at VRVis?
Hesina: We have only recently experienced something of a ‘shortage’ because the demand for our solutions has risen sharply. I could hire five new employees with well-paid contracts tomorrow. I would also say that, in contrast to the US, there is a lack of both successful lighthouse projects and digital shining lights in this country. That’s why IT jobs are often considered somewhat unattractive here. But this reputation is completely unfounded and unjustified. The nerd image from the past may still have an effect here. An image campaign would do the entire industry good. This could start in the lower grades of secondary schools.

Would you say that VRVis enhances Vienna as a business location?
Hesina: Definitely – on the one hand, because as a software company we employ a relatively high number of staff in the research area. On the other hand, because as a COMET centre we are also increasingly active in the SME sector. There, we provide an important transfer service from research to business and industry, which helps to drive the digital transformation and make the business location more attractive through innovations.

What is the next trend in AI?
Hesina: In the public eye, AI is definitely first of all continuing to be about connecting textbased AI with search functions and the like. We are primarily focusing on optimising AI and machine learning for image and pattern recognition and making them more trustworthy and robust through customised visualisations – from manufacturing to medicine, our solutions are already indispensable assets that will become much more important in the future.