Research Seminar on StyleGAN-based 3D Stylized Face Generator & LLM based XR Applications

Research Seminar with Prof. Seungyong Lee and Dr. Chang-Hun Kim: Toonify3D: StyleGAN-based 3D Stylized Face Generator and GPT + XR: LLM based XR Applications

We are pleased to invite you to a research seminar titled “Toonify3D: StyleGAN-based 3D Stylized Face Generator and GPT + XR: LLM based XR Applications”, presented by Prof.Seungyong Lee, professor of POSTECH, Korea and Dr. Chang-Hun Kim, CEO of DeepXRLab

Location: Room I 201, VinUniversity

Date and Time: 16:00 – 17:30, June 12, 2025 (Thursday)

Expected attendance: 30 participants

About the Speaker

Prof. Seungyong Lee is a professor of Computer Science and Engineering at POSTECH, Korea. He obtained his Ph.D. from KAIST in 1995, and conducted postdoctoral research at City College of New York. Since 1996, he has led the Computer Graphics Group at POSTECH. His sabbatical appointments include MPI Informatik and Adobe Systems, where his research on image deblurring and upright adjustment was integrated into Adobe Creative Cloud. His current research focuses on 3D modeling, human avatars, and computational photography.

Dr. Chang-Hun Kim is the CEO of DeepXRLab and an Emeritus Professor at Korea University. He received his Ph.D. in Electronics and Information Science from the University of Tsukuba in 1993. His research interests include AI, extended reality (XR), and physics-based simulation. Throughout his 30-year career, he has published extensively in prestigious venues such as SIGGRAPH, Eurographics, and IEEE TVCG. He previously served as Department Chair and Dean at Korea University and as President of the Korea Computer Graphics Society. Currently, he leads DeepXRLab’s efforts in advancing generative AI and XR technologies.

Seminar Overview
Recent advancements in AI and immersive media are transforming human-computer interaction, particularly through the integration of Large Language Models (LLMs), Extended Reality (XR), and 3D generative technologies. This seminar explores cutting-edge applications at the intersection of AI, language, and visual computing.

The first presentation from DeepXRLab introduces AI-powered virtual characters that can engage in real-time conversations, recognize emotions, and express themselves through speech and facial animations. These characters are applied in healthcare and business contexts, such as personalized wellness coaching and interactive medical avatars. The talk will also discuss domain-specific LLMs and the design of engaging, safe AI agents.

The second presentation focuses on Toonify3D, a state-of-the-art framework that transforms 2D facial stylization into expressive 3D avatars. Leveraging the strengths of StyleGAN and novel 3D-lifting techniques (StyleNormal), Toonify3D enables full-head 3D avatar generation and dynamic facial editing, with applications ranging from entertainment to virtual communication.

Together, these sessions showcase how AI-driven language, vision, and 3D modeling are shaping the future of immersive, emotionally intelligent digital experiences.

For Inquiries (QR code below)