Generating high-quality 3D visual data, such as 3D models that can be integrated into interactive virtual environments or fabricated as real-world objects via 3D printing, remains a core challenge in generative AI. Existing 3D generative models are largely constrained to producing static scenes, lacking the ability to capture complex motions or simulate dynamic object interactions. In this talk, I will first present GEOPARD, a transformer-based architecture that infers articulation and motion attributes from a single static snapshot of a 3D shape. Then, I will discuss SOPHY, a new deep generative model that jointly synthesizes object shape, texture, and physics-aware material properties, enabling the creation of 3D assets that are ready for simulation in dynamic, interactive environments. I will also discuss a number of applications, including text-to-4D generation, 3D reconstruction of physically plausible shapes from a single image, and image synthesis from 3D guidance and text prompts. This work pushes the boundaries of generative modeling from static 3D to dynamic, physically grounded 4D content, opening up new possibilities for virtual worlds, simulation, and design automation.
Evangelos Kalogerakis' research deals with the development of graphics+vision algorithms and techniques, empowered by Machine Learning and Artificial Intelligence, to help people to easily create and process representations of the 3D visual world. He is particularly interested in algorithms that generate 3D models of objects, scenes, animations, and intelligently process 3D scans, geometric data, collections of shapes, images, and video. His research has been supported by the European Research Council (ERC consolidator grant) and grants from the National Science Foundation (NSF). He is currently an Associate Professor at the School of Electrical and Computer Engineering at the Technical University of Crete, where, starting in 2025, he leads a research group focused on graphics and vision. Previously, he was a tenured Associate Professor at the College of Information and Computer Sciences at the University of Massachusetts Amherst, which he initially joined as an Assistant Professor in 2012. Before that, he was a postdoctoral researcher at Stanford University from 2010 to 2012. He earned his PhD from the University of Toronto in 2010. His PhD thesis introduced machine learning techniques for geometry processing. He has served as Area Chair in CVPR, ICCV, ECCV, NIPS and on technical paper committees for SIGGRAPH, SIGGRAPH ASIA, Eurographics, and the Symposium on Geometry Processing. He has also served as an Associate Editor in the Editorial Boards of IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) and IEEE Transactions on Visualization & Computer Graphics (TVCG). He co-chaired Eurographics 2024. He was listed as one of the 100 most cited computer graphics scholars in the world between 2010-2020 by the Tsinghua's AMiner academic network.
