Explore the groundbreaking work of Prof. LIU Yuan, Assistant Professor of the Division of Integrative Systems and Design (ISD), whose research merges AI with 3D technology to revolutionize how we create and experience digital worlds! From advanced 3D reconstruction to generative modeling, his innovations are transforming industries from gaming to manufacturing while making spatial intelligence more accessible to everyone.

-----

Could you tell us more about you e.g. where do you from and why joining HKUST, and something you are working on now e.g. current research, and course that you will teach in coming Spring term?

Before joining HKUST, I earned my PhD from HKU and conducted my postdoctoral research at NTU, Singapore. My expertise lies in computer graphics and 3D computer vision, with a strong passion for pushing the boundaries of AI-driven 3D content creation—including 3D reconstruction, generative 3D modelling, 3D design, and 3D-aware video generation.

At the core of my research is the fusion of AI and 3D technology to digitally recreate and generate the world around us. 3D reconstruction plays a crucial role in various applications, from building digital replicas of real-world environments for localization and navigation to creating detailed 3D maps that aid both humans and robots in planning and interaction. Meanwhile, 3D generation powers the creation of virtual assets for industries such as gaming, film, and manufacturing, allowing us to bring entirely new worlds to life.

I’m thrilled to be part of the Division of Integrative Systems and Design (ISD) within the Academy of Interdisciplinary Studies (AIS) at HKUST. ISD’s interdisciplinary environment—spanning fields like materials science, robotics, and design—provides an incredible opportunity to collaborate across disciplines. By integrating my 3D research with diverse fields, I hope to contribute to innovative applications that enhance both technology and daily life.
 
What inspired you to specialize in this line of research?
I believe that understanding and generating the world in 3D—and even in 4D by incorporating the dimension of time—will be a cornerstone of future AI technologies. After all, we humans experience and interact with our surroundings in 3D, making spatial awareness essential for truly intelligent systems. That’s why my research focuses on reconstructing the 3D world and advancing AI’s ability to create new 3D content. Ultimately, I envision this work leading to the emergence of 3D spatial intelligence—AI that perceives, understands, and generates the world as naturally as we do.

What impact do you want your work to have on society?

My primary goal is to empower professional designers in manufacturing, video game development, and film production by streamlining the 3D modelling process. By reducing the need for tedious, repetitive tasks, my research allows creators to focus on what truly matters—the artistic and imaginative aspects of their work. Looking ahead, I hope my work will drive the widespread adoption of 3D content in everyday life, making high-quality 3D creation accessible to everyone, even amateur users. Imagine a future where anyone can effortlessly generate detailed 3D models and volumetric videos. For instance, instead of simply sharing photos or videos of a new home, we could send friends a fully interactive 3D model, allowing them to explore the space as if they were there in person.

Do you have any advice for students interested in your research area?  
Beyond mastering fundamental AI techniques such as models and loss functions, diving into 3D AI content creation (3D-AIGC) requires a strong foundation in multiview geometry and computer graphics. Multiview geometry provides the mathematical backbone for reconstructing the 3D world from 2D images, covering key concepts like camera models and epipolar geometry. Meanwhile, computer graphics encompasses crucial topics such as geometry processing, rendering pipelines, and animation—powerful tools that bring 3D research to life. By combining AI with these disciplines, we can push the boundaries of 3D content generation and revolutionize the way digital worlds are created.

Could you share a fun fact about you?
I've been a passionate gamer for as long as I can remember, from the classic The Legend of Sword and Fairy (仙劍奇俠傳) to DoTA, League of Legends, Genshin Impact, and Honkai: Star Rail. The rich stories and immersive worlds in these games have always captivated me, sparking my curiosity about how such experiences are crafted. This fascination has driven me to develop AI tools that can revolutionize video game design, making it easier to create breathtaking worlds and compelling narratives.