A promotional film for a joint mathematics master's program between the University of Konstanz and the University of Bern, utilizing generative AI to visualize abstract concepts.
To promote this joint double-degree program, I used generative AI to transform static university photography into cinematic sequences and visualize abstract mathematical equations.
By expanding existing imagery and generating new assets where photoshoots were impossible, I established a clean, technical aesthetic that reflects the program's academic rigor.
I handled the full-cycle production, including the conceptual design, motion graphics, and original music composition.
It was about finding the beauty of mathematics and treating the lack of footage as a creative challenge. The goal was to produce something accurate yet interesting enough to encourage people to research the program.
Since I couldn't access the university buildings, I had to find a way to convey the image of mathematics not just through equations, but visually, using a very limited amount of existing material.
The soundtrack rooted my thinking. As I conceptualized the script, I used the music to set the tone while I played with Midjourney to create abstract images of geometries, stochastics, and topologies.
This allowed me to start a conversation with the university teams about what made sense to visualize and how to balance the abstract with the academic.
There was no way to fool this audience; they are well-versed in the subject. We solved the accuracy issue by introducing actual equations and 3D-tracking them into the shots using Adobe After Effects.
By making the equations appear as if they were floating in space or rooted in the architecture, we bridged the gap between dry academic rigor and a cinematic aesthetic.
I had to abandon many abstract ideas because the AI-generated visuals didn't make sense mathematically or visually. We went through many iterations to find the right balance.
Initially, there was more abstraction, but we eventually toned it down. We shifted toward using actual university photography enhanced by AI rather than purely synthetic images, ensuring the spaces remained recognizable.
Aerial and architectural photography without people worked best; the AI rendered those very accurately. One of the biggest technical hurdles was taking vertical images from the university libraries and expanding them into a 16:9 cinematic ratio.
While the AI struggled with human figures, it was almost perfect at architectural visualization, allowing me to create cinematic sequences from static photos.
See other projects
