Zirui Wu

Zirui Wu received his B.Sc. degree in Computer Science and Technology from Beijing Institute of Technology, China.

He is currently pursuing his Ph.D. degree in robotics and autonomous systems (ROAS) at the Hong Kong University of Science and Technology (Guangzhou), co-supervised by Prof. Martin R. Oswald (the University of Amsterdam) and Prof. Jie Song.

During 2024-2025, he was a reserach intern at NIO, worked with Kang Xu on specular neural inverse rendering in driving scenarios.

Email  /  CV  /  Google Scholar  /  Github

Research

I'm interested in 3D perception for robotic applications. Representative papers are highlighted.

We propose using point cloud projections instead of Plücker coordinates as inputs for feed-forward view synthesis. This enables consistent rendering and MAE-style self-supervised learning.

We present DualMap, a novel online open-vocabulary mapping system that enables robots to understand and navigate dynamically changing environments through natural language queries.

We propose a real-time approximated global illumination method for 3D Gaussian inverse rendering. The key idea is to leverage indirect emissors in the screen space to approximate the global illumination effect.

We propose an autonomous driving simulator based upon neural radiance fields (NeRFs). Compared with existing works, ours has three notable features: (1) Instance-aware. (2) Modular. (3) Realistic. Our simulator achieves state-of-the-art photorealism results given the best module selection. Our simulator will be open-sourced while most of our counterparts are not.

We propose a two-stage pipeline for global localization in a city-scale neural map.

Web page borrowed from Jonathan T. Barron. Last update: .