FIT BUT welcomes a new professor. Martin Čadík has taken up his appointment

On Tuesday, December 16, 2025, in the Great Hall of Prague's Karolinum, President Petr Pavel, in the presence of Minister of Education Robert Plaga, presented a total of 69 appointment decrees to new professors. The decree was also received by the new professor of the Faculty of Information Technology in the field of Computer Science and Informatics. Martin Čadík (Department of Computer Graphics and Multimedia) has been working at our faculty since 2013; two years later, he became an associate professor, and now, after ten years, he has added the highest scientific and pedagogical title.

Čadík is the head of the research group CPhoto@FIT, which, as the name suggests, deals with the connections between computational photography. Its research methods lie at the intersection of image processing, computer vision, graphics, physics, visual perception, and other fields. The group focuses on HDR image and video processing, image and video quality assessment, computational aesthetics, and human visual perception.

Geolocation: When a photo can say more

Martin Čadík's main area of interest is geolocation, i.e., the use of geographical or topographical models of the planet's surface to process digital photographs in order to obtain additional information and create a new layer of information that expands on the original image. Typically, this involves determining the position or orientation of the camera, where the photographer was standing when the pictures were taken, the name of the mountain captured in the photograph, and so on. Čadík has been involved in computational photography and geolocation for over 20 years. In doing so, he partly combines scientific interest with personal interest. "I enjoy mountains and nature, and I often do research with outdoor photos. And they are often my own photos," he comments on the image, which he pairs with the model during the interview. "I get a photo, generally knowing nothing about it, and my goal is to find out where it was taken and which direction the photographer was looking. We apply our algorithm to this, and the result is the alignment of the photo with the rendered model data using the correspondences found." The model itself is a carrier of new information, allowing a number of things to be discovered from a single image. This includes botanical (e.g., vegetation) and geographical or topographical data (identification of hills, altitude, etc.). This opens up further possibilities for editing and presenting digital photos, for example in the form of augmented reality.

Martin Čadík and his colleagues use machine learning techniques, which, as Čadík points out, are now commonly referred to as AI. Historically, these techniques have been closely related to the field of computer vision. "From today's perspective, we can say that we have always been involved in AI computer vision. Currently, however, the term is used very broadly." He himself uses, for example, transformer-type neural networks or convolutional neural networks, for which he collects rendered data for training.

When asked about specific applications of the created algorithm and methods, Martin Čadík mentions a number of items. For example, data from the model precisely defines the distances between objects in the shot, which can be measured. This makes it possible to change the depth of field or simulate shots from huge lenses that would otherwise be impossible to produce. A specific area of application is augmented reality, which, for example, supplements the image in the camera of the user's mobile phone with additional information, or, during subsequent processing on a computer, allows for image adjustments that would otherwise be very complicated. Users can refocus on another object, add shadows, or change the lighting of a photo. The software can also place the captured photo directly in the terrain and then return the photographer to the place where the photo was taken in virtual reality. Thanks to special glasses, the photographer can then show others the place where he took the photo and what the surroundings outside the photo look like by "flying" through a model supplemented with images. The software was developed by Čadík's team in collaboration with Adobe Research, and the results have already been patented. And geolocation naturally also has security and forensic applications, Čadík points out.

However, Čadík's scientific career is by no means focused on just one topic. He himself mentions HDR technology, which he worked on during his time at the Max Planck Institute for Informatics. He also mentions the extremely broad scientific challenge of human visual perception. He has also gained professional insight into subfields such as refotography.

Commitment and the future

Martin Čadík sees his new professorship as a commitment: "I appreciate it. I think the time was right for a professorship. It brings to mind the growing responsibility that comes with it, if only because a professor's statements carry more weight in the public sphere, which is why maximum accuracy is required. On the other hand, I know a number of experts who are excellent but do not yet have a professorship or are not actively seeking one." He also mentions the aspect of educating future talent: "It is not only a scientific but also a pedagogical title. I feel a strong commitment to passing on my experience to students and doctoral candidates."

Martin Čadík is quite clear about future professional challenges: "We are constantly improving methods of computer vision burdened by a lot of noise. We are working on improving methods on several fronts. The goal is more accurate location identification." He also mentions the possibility of detecting photographic forgeries or manipulations (in the sense of post-editing) of photographs. He would also like to continue participating in activities within his professional community. For many years, he has been organizing meetings of Czech and Slovak colleagues, HiVisComp. The eleventh edition of the event will take place in January 2026 in the High Tatras.

We congratulate Martin Čadík on his professorship and wish him continued enjoyment in his research, as is evident from this short interview with him.