Kindrat Beregovyi
Ph.D. Student
Computer Science
Thursday, June 12, 2025, 11:00 a.m.
Chase 130
Abstract
This presentation will highlight the development of Focal Engine, a specialized visualization engine designed to bridge the gap between game engines and scientific visualization tools. It will focus on both completed contributions enabled by the engine, and future improvements that will enable further novel research applications. One such contribution is research on AI upscaling techniques such as DLSS/DLAA and their impact on the accuracy of scientific visualization.
The talk will demonstrate existing projects that showcase the engine's capabilities alongside planned enhancements and future applications. The project list includes HabiCAT 3D, an application for coral reef complexity analysis that supports models with up to 300 million triangles, and calculates new 3D complexity metrics. Another project is a VR point cloud editor capable of working with 400+ million points, functionality that no widely-used commercial engine can accomplish out-of-the-box. Additionally, Focal Engine powers a unique VR client for cloud-based bathymetry processing.
Finally, the ongoing development of volumetric rendering is illustrated through visualization of NOAA's STOFS-3D ocean forecast model output. Volumetric analysis and rendering will also be used in HabiCAT 3D to expand its capabilities.
Bio
Kindrat Beregovyi is a computer science Ph.D. student at UNH. In Ukraine, he worked as a business application programmer and later had experience as a junior graphics engineer programmer. As a research assistant in CCOM, he will be working on different visualization problems. If you are interested in computer graphics and 3d engines in particular and want to discus these topics, feel free to contact him.
