Report Title: High-Fidelity Virtual Digital Humans: From Modeling, Driving to Editing and Generation

Time: 14:40, Saturday, November 1, 2025

Venue: Room 1602, Building A, Science and Education Building, Feicui Lake Campus

Speaker: Professor Juyong Zhang

Affiliation: University of Science and Technology of China

Organizer: School of Computer Science and Information Engineering

Report Abstract:

In recent years, differentiable radiance field representations represented by 3D Gaussian Splatting (3DGS) have achieved tremendous breakthroughs in 3D reconstruction of humans, objects and scenes due to their powerful fitting capability and differentiability. For large-scale application of virtual digital humans, in addition to high-fidelity rendering, multimodal driving, as well as appearance and emotion editing and generation of digital humans are also crucial. In this report, I will introduce the research work of our group on efficient reconstruction of high-fidelity virtual digital humans from convenient inputs such as monocular videos or single images, physics-based character and clothing driving, and multimodal editing and generation.

Speaker Profile:

Juyong Zhang is a Professor at the School of Mathematical Sciences, University of Science and Technology of China (USTC), and Vice Dean of the USTC Institute of Advanced Technology. He is supported by the National Natural Science Foundation of China for Excellent Young Scholars and the Outstanding Member of the Youth Innovation Promotion Association of the Chinese Academy of Sciences. He serves on the Editorial Boards of IEEE Transactions on Mobile Computing and IEEE Computer Graphics and Applications. His research interests include computer graphics and 3D vision, focusing on efficient and high-fidelity 3D digitization of the real physical world and creation of high-realistic virtual digital content.