Courses are taught by internationally known experts. Instructors typically include an overview of the state of the art of their topic, and highlight their own research, but also include the current work of others. It is intended to be a “course” in the traditional sense of enabling attendees to go forth and produce new results of their own, rather than simply use existing knowledge.
- Robust Low-Poly Meshing for Mobile Games
Dr. Xifeng Gao, Principal Researcher at LightSpeed Studios - 40+ Years of Subdivision Schemes
Prof. Lucia Romani, University of Bologna - Efficient High-Quality Surface Meshing with Anisotropy
Prof. Zichun Zhong, Wayne State University
Robust Low-Poly Meshing for Mobile Games
Dr. Xifeng Gao
Abstract: Low-poly meshes are essential in the development of mobile games due to their efficiency in real-time rendering, animation, and data storage. However, creating high-quality low-poly models often requires manual effort to strike the ideal balance between a minimal polygon count and the preservation of visual and shape fidelity, among other considerations. This process can be time-consuming, even for skilled artists, and typically involves a tedious cycle of trial and error.
In this talk, I will introduce several techniques that leverage domain knowledge in geometry processing, computer graphics, and machine learning to automatically generate low-poly meshes optimized for mobile game applications.
Biography: Xifeng Gao is a principal researcher at LightSpeed Studios. He has more than 10 years of academic research experience, with a focus on geometry computing. The target applications are computer graphics, computer-aided-design and analysis, and robotics. He has more than 60 peer-reviewed papers published in top-ranked journals and conferences. More information can be found on https://gaoxifeng.github.io/.
40+ Years of Subdivision Schemes
Prof. Lucia Romani
Abstract: Subdivision schemes are iterative methods for generating smooth, highly detailed surfaces from coarse polygonal meshes of arbitrary topology. Due to their efficiency and flexibility, in late 1990s they started to achieve widespread use in computer graphics and animation, particularly for character modeling. More recently, it was also shown that subdivision surfaces can be advantageously exploited for segmenting volumetric objects in 3D biomedical images. However, the use of subdivision schemes has not yet expanded to sectors like high-end computer-aided design in the automotive industry, as the quality of the resulting surfaces does not yet meet the required standards. In this talk, we will review the major advances in the field since the groundbreaking work by Catmull and Clark (1978) which generalized bi-cubic uniform B-spline surfaces to quadrilateral meshes with arbitrary topology. Specifically, we will focus on strategies introduced over the past two decades to overcome the main limitations of Catmull-Clark subdivision surfaces. The primary goal will be, on the one hand, to show how smoothness and curvature behavior can be enhanced and, on the other hand, to illustrate how challenges such as shape control through free parameters and the generation of specific geometries can be addressed.
Biography: Lucia Romani received a PhD in Computational Mathematics (research area of Numerical Analysis) from the University of Padua (Italy) in 2004. After working for seven years as an assistant professor at the Department of Mathematics of the University of Milan-Bicocca (Italy), in 2014 she was promoted to associate professor. In 2018 she moved to the Department of Mathematics, University of Bologna (Italy) where she was promoted to full professor in 2022.
Her research interests span the fields of approximation theory, geometric modelling and CAGD, with a special focus on splines, subdivision schemes and their applications.
Lucia is author and co-author of 80+ peer-reviewed scientific publications, 70+ in international journals.
More information can be found on https://www.unibo.it/sitoweb/lucia.romani/en.
Efficient High-Quality Surface Meshing with Anisotropy
Prof. Zichun Zhong
Abstract: In geometric modeling, physical simulation, and mechanical engineering fields, anisotropic meshes are crucial for performing better shape approximations and achieving higher accuracy in numerical simulations. Anisotropic meshes are designed with elongated mesh elements that follow prescribed directions and stretchings. With the same number of budgeting elements, the anisotropic meshing representation for the surface of the original model is preferred due to the smaller approximation errors compared with the correspondent isotropic counterpart. However, the primary bottleneck in anisotropic mesh generation lies in its computational complexity and time-intensive nature. In this short course, I will introduce a streamlined set of techniques from traditional optimization-based methods to newly learning-based methods to efficiently generate high-quality and high-fidelity anisotropic surface meshes.
Biography: Zichun Zhong is an Associate Professor of Computer Science at Wayne State University (WSU). He received a Ph.D. degree in Computer Science at The University of Texas at Dallas in 2014 and was a postdoctoral fellow in the Department of Radiation Oncology at UT Southwestern Medical Center at Dallas from 2014 to 2015. His research interests focus on geometric modeling and computing for 3D computer graphics, computer vision, visualization, and medical image processing. He has published over 60 peer-reviewed conference and journal papers in his research fields. He is associate editor of GMOD, guest editor of CAGD and C&G. He has served as a Technical Paper Chair of SMI 2024, 2025, and a Program Co-Chair of GMP 2024. He has served as a program committee member and area chair for many international conferences. He received the NSF CAREER Award in 2019, NSF CRII Award in 2016, COE Faculty Research Excellence Award at WSU in 2019, COE Award for Excellence in Teaching at WSU in 2020. More information can be found on https://zichunzhong.github.io/.
Contact details
For any question concerning the short courses, please contact the chair:
- Xiaohu Guo, Short Courses Chair
The University of Texas at Dallas. email: xguo@utdallas.edu
