Xiao Fu (付潇)

Ph.D @ MMLab, CUHK

Hi, stranger, welcome to stand by! I am a Ph.D. at Multimedia Laboratory (MMLab) of The Chinese University of Hong Kong (2023-Present), supervised by Prof. Dahua Lin. I received my B.Eng (Honors) degree (2018-2022) at Zhejiang University, supervised by Prof. Yiyi Liao. I was lucky to have research collaboration with Prof. Andreas Geiger. I am dedicated to enhancing my professional skills in the field of machine learning and cultivating innovation ability.

Research interest: compact and efficient visual computing, including

  • Foundation GenAI (Image/Video/3D/4D)
  • Visual Reconstruction, Analysis and Reasoning
  • User-friendly Controlbility and Editing

News

Mar. 15, 2025 I will intern at Deep Imagination Research @ NVIDIA Research. See you in Santa Clara!
Mar. 14, 2025 Check our 3D-aware video generation works (3DTrajMaster, SynCamMaster, and ReCamMaster)
Dec. 30, 2024 Invited talk at AnySyn3D Webinar on ''3D Interactive Video Generation''.
Jul. 01, 2024 One paper is accepted to ECCV 2024
Oct. 20, 2023 Invited talk at Shenlan Open Courses on ''Panoramic Label Rendering for Autonomous Driving Simulator''.
Aug. 01, 2023 Co-orgranize OmniObject3D Challenge in AI for 3D Content Creation Workshop at ICCV'23.
Mar. 31, 2023 Awarded Hong Kong PhD Fellowship Scheme (HKPFS). Thx study life in ZJU!
Feb. 28, 2023 One paper is accepted to CVPR 2023, as Best Paper Award Candidate
Aug. 02, 2022 One paper is accepted to 3DV 2022

Selected Research

Full publication list can be found on Google Scholar

  1. 3DTrajMaster: Mastering 3D Trajectory for Multi-Entity Motion in Video Generation
    3DTrajMaster: Mastering 3D Trajectory for Multi-Entity Motion in Video Generation
  1. GeoWizard: Unleashing the Diffusion Priors for 3D Geometry Estimation from a Single Image
    GeoWizard: Unleashing Diffusion Prior for 3D Geometry Estimation from a Single Image
  1. Panoptic Neural Representation for 360º 3D-to-2D Label Transfer in Urban Scenes
    PanopticNeRF-360: Panoramic 3D-to-2D Label Transfer in Urban Scenes
    Xiao Fu, Tianrun Chen, Yichong Lu, Xiaowei Zhou, Andreas Geiger, and Yiyi Liao
    1. Panoptic Neural Representation for 3D-to-2D Label Transfer in Urban Scenes
      Panoptic NeRF: 3D-to-2D Label Transfer for Panoptic Urban Scene Segmentation
    1. ReCamMaster: Camera-Controlled Generative Rendering from A Single Video
      ReCamMaster: Camera-Controlled Generative Rendering from A Single Video
    1. OmniObject3D: Large-Vocabulary 3D Object Dataset for Realistic Perception, Reconstruction and Generation
      OmniObject3D: Large-Vocabulary 3D Object Dataset for Realistic Perception, Reconstruction and Generation
      Tong Wu, Jiarui Zhang, Xiao Fu, Yuxin Wang, Jiawei Ren, Liang Pan, Wayne Wu, Lei Yang, Jiaqi Wang, Chen Qian, Dahua Lin, and Ziwei Liu

Experiences

Research Intern, Deep Imagination Research, NVIDIA Research
Santa Clara, United States
Topic: Video Foundation Model
Advisor: Ming-Yu Liu
Research Intern, Kuaishou Kling
Shenzhen, China
Topic: Controllable Video Generation
Advisor: Xintao Wang
Associate Researcher, Shanghai AI Innovation Center
Shanghai, China
Topic: 3D Reconstruction and Generation
Advisor: Dahua Lin

Honors & Awards

CVPR Best Paper Award Candidate 2023
Hong Kong PhD Fellowship Scheme (HKPFS), Hong Kong SAR 2023
CUHK Vice-Chancellor HKPFS Scholarship 2023
Outstanding Graduation Thesis Award of Zhejiang University 2022
National Scholarship, Ministry of Education of P.R. China 2020, 2021
ZJU ISEE Excellent Student Award, Dean's Honor Graduate 2021
ZJU ISEE Yuelun Alumni Scholarship, Dean’s Award for Academic Contest 2020, 2021
Gold Prize, 7th “Internet Plus” College Student Innovation and Entrepreneurship Contest, Team Leader 2021
Finalist Winner, Mathematical Contest in Modeling (MCM/ICM) 2020