Ciril Bohak

An assistant professor at the University of Ljubljana, Faculty of Computer and Information Science, and a member of the Laboratory for Computer Graphics and Multimedia. Currently, he holds the position of president of the IEEE Slovenia Computer Society and is among the founding members of the Slovenian Human-Computer Interaction Society. His primary research areas include Computer Graphics, focusing on volume rendering, physically-based rendering, neural rendering, and differentiable rendering; Data Analysis and Visualization, with a focus on biological, medical, high-energy physics, geodetic, various geo-referenced and multimedia data; Remote Sensing, emphasizing satellite/airplane image segmentation/analysis; and Human-Computer Interaction, particularly regarding novel use of 3D and VR and related interactive user studies. Throughout his career, he has contributed to various projects funded by the EU, the Slovenian National Research Agency, and industry partners, connecting research discoveries to practical implementations.

PostDoc Topics:

Data Analysis and Visualization of High-Energy Physics

Experiments in high-energy physics produce an immense volume of data that necessitates thorough analysis and visualization. With the planned upgrades to the existing research infrastructure, the volume of data will increase even further. Analyzing this data is crucial for researchers to decode the underlying phenomena, helping them draw conclusions about the insights, evidence, or contradictions the data presents. The main goal of this project is to develop methods, both semi-automatic and automatic, for analyzing this data in a visualization framework. We aim to efficiently summarize, filter, and organize the data, ensuring it is both accessible and meaningful for users. A significant amount of the data has already been analyzed, and the insights gained from this previous work will lay the groundwork for creating new, advanced techniques. These new methods, which will include both semi-supervised and unsupervised automated approaches, are designed to generate insightful visualizations. They will capitalize on the latest developments in visual analytics and the broader field of visualization to enhance our understanding of the data.

Procedural modeling of internal cellular structure based on real-world data

Thanks to advanced data acquisition techniques, we now have the ability to collect vast quantities of data that provide deep insights into the inner structure of cells. This analysis, which looks into the number and distribution of cellular components, is key to unraveling how cells operate and the processes they undergo. To support such studies, having a substantial volume of segmented and classified data for analysis is invaluable. This data allows us to uncover detailed information on the architecture of cellular organelles, their placement within cells, and how different organelle types are spatially related to one another. Such insights pave the way for a more profound understanding of the precise locations, mechanisms, and timings of specific cellular processes, including the structures involved and the necessary conditions for these processes to occur. However, one major challenge in this approach is the task of data labeling. Our aim is to develop methods that can automatically extract this information without manual supervision, utilizing it in the generation of comprehensive cell models through generative modeling techniques.