Wei Hu (胡玮)

Associate Professor, Ph.D. Supervisor

Peking University Boya Young Fellow

Wangxuan Institute of Computer Technology

Peking University, Beijing, China

IEEE Senior Member

Email: forhuwei AT pku DOT edu DOT cn


I am a tenured associate professsor and independent PI leading the GLab at Wangxuan Institute of Computer Technology, Peking University. I obtained the B.S. degree in Electrical Engineering from University of Science and Technology of China in 2010, and the PhD degree in Electronic and Computer Engineering from The Hong Kong University of Science and Technology in 2015. Before joining Peking University, I was a Researcher in the Imaging Science Laboratory of Technicolor, Rennes, France. Besides, I used to be a visiting student at National Institute of Informatics, Japan, supervised by Prof. Gene Cheung. I have collaborations with Prof. Antonio Ortega from the University of Southern California, Prof. Xin Li from West Virginia University, etc.  

My research interests include Graph Signal Processing, Graph-based Machine Learning and their applications in the processing, analysis and synthesis of geometric data and beyond (structural data such as images, network data, brain signals, etc.), which lies at the intersection of signal processing and machine learning. You may click the "Research" tab to know more about my research interests.

Graph Signal Processing
(image credit: Catherine Amein)		 Graph-based Machine Learning 3D Visual Computing

 

Positions Opening: We are looking for self-motivated postdocs, graduate students and interns. If you are passionate about our research on Graph Signal Processing and Graph-based Machine Learning with applications in 3D visual computing, please send a detailed CV to me.

News

  • 11/10/2025: 1 AAAI paper accepted

  • 09/19/2025: 1 NeurIPS paper accepted (spotlight)

  • 07/25/2025: 1 TNNLS paper accepted

  • 07/09/2025: 1IJCV paper accepted

  • 07/05/2025: 2 ACM MM paper accepted

  • 06/27/2025: 1 TNNLS paper accepted

  • 06/26/2025: 2 ICCV paper accepted

  • 06/23/2025: 博士生刘岱宗获评北大优秀博士学位论文。Daizong Liu receives PKU Outstanding Doctoral Dissertation Award.

  • 05/30/2025: 1 ICANN paper accepted

  • 05/20/2025: elected as the ICME 2026 TPC Co-Chair representing the MMSP-TC

  • 05/18/2025: 1 TACD paper accepted

  • 03/31/2025: 1 ICME paper accepted

  • 02/27/2025: 1 CVPR paper accepted

  • 02/18/2025: 1 TOMM paper accepted

  • 01/26/2025: 1 APSIPA paper accepted

  • 08/18/2024: 1 ICPR paper accepted

  • 07/16/2024: 2 ACM MM paper accepted

  • 07/01/2024: 1 ECCV paper accepted

  • 03/15/2024: Our work on system-level time computation and representation in the suprachiasmatic nucleu in collaboration with Prof. Heping Cheng has been accepted to Cell Research!

  • 02/27/2024: 1 CVPR paper accepted

  • 11/28/2023: 1 TPAMI paper accepted

  • 07/14/2023: 1 ICCV paper accepted

  • 02/28/2023: 1 CVPR paper accepted

  • 01/16/2023: 1 TMM paper accepted

  • 09/30/2022: congratulate Daizong Liu for the National Scholarship

  • 08/19/2022: invited to give a talk at CCIG 2022 [slides]

  • 07/21/2022: 1 TPAMI paper accepted

  • 07/04/2022: 2 ECCV papers accepted

  • 06/30/2022: 2 ACM MM papers accepted

  • 06/17/2022: congratulate Bi'an Du for "Excellent Undergraduate" both in Peking University & Beijing

  • 06/17/2022: congratulate Bi'an Du for the Top-10 Excellent Undergraduate Thesis Award in EECS, Peking University

  • 04/27/2022: 1 TPAMI paper accepted

  • 01/12/2022: our ICIP Special Sesstion Proposal "Point cloud compression and processing" has been accepted

  • 12/11/2021: our ICME Special Session proposal on "Advances in Point Cloud Acquisition, Processing and Understanding" has been accepted

  • 12/01/2021: 1 TKDE paper accepted

  • 11/23/2021: 1 TIP paper accepted

  • 11/03/2021: Elected as a member of IEEE Multimedia Signal Processing Technical Committee (MMSP-TC)

  • 09/04/2021: 1 TMM overview paper on "Graph Signal Processing for Geometric Data and Beyond: Theory and Applications" accepted

  • 08/31/2021: 1 TIP paper accepted

  • 08/02/2021: invited to serve as tutorial co-chair for VCIP 2022

  • 07/24/2021: attend APSIPA panel discussion on "Future of Graph Signal Processing”as a panelist

  • 07/23/2021: 1 ICCV paper accepted

  • 07/08/2021: Received the 2021 IEEE Multimedia Rising Star Award-Honorable Mention for "outstanding early-stage career achievements in the area of geometric data processing in the graph domain"

  • 07/08/2021: Received ICME 21' Outstanding Service Award

  • 07/04/2021: 1 ACM MM paper accepted

  • 06/21/2021: 1 TIP paper accepted

  • 06/21/2021: 1 TPAMI paper accepted

  • 06/21/2021: congratulate Shitong Luo for the Top-10 Excellent Undergraduate Thesis Award in EECS, Peking University

  • 06/14/2021: our paper "Diffusion Probabilistic Models for 3D Point Cloud Generation" has been selected as CVPR 2021 Best Paper Candidate

  • 06/06/2021: our paper won Best Poster Award in The CAAI International Conference on Artificial Intelligence

  • 05/04/2021: Appointed as Associate Editor for Signal Processing Magazine

  • 04/20/2021: Elevated to the grade of IEEE Senior member

  • 04/15/2021: Appointed as Associate Editor for Transactions on Signal and Information Processing over Networks

  • 04/09/2021: Our ICCV workshop "When Graph Signal Processing meets Computer Vision" is accepted [link] [PDF]

  • 03/24/2021: Appointed as Associate Editor for Frontiers in Signal Processing

  • 03/01/2021: 2 CVPR papers accepted

  • 02/15/2021: congratulate Shitong Luo who receives an offer from CMU

More

  • 12/29/2020: 1 TMM paper accepted

  • 12/05/2020: 1 INFOCOM paper accepted

  • 11/24/2020: Appointed Associate Member in MMSP-TC in SPS

  • 11/03/2020: selected as Peking University Boya Young Fellow

  • 08/13/2020: elected as MSA-TC Member (September 1, 2020 to August 31, 2024)

  • 08/03/2020: 1 TCSVT paper accepted

  • 07/26/2020: 1 ACM MM paper accepted

  • 07/10/2020: our paper "3D Dynamic Point Cloud Inpainting via Temporal Consistency on Graphs" won Best Student Paper Runner Up Award in ICME 2020

  • 07/03/2020: 1 ECCV paper accepted

  • 06/23/2020: serve as an Open Source Chair for ICME 2021

  • 06/22/2020: Xiang Gao won President Scholarship, Peking University (2020-2021)

  • 05/30/2020: congratulations for students who receive offers from Cornell Tech, USC, JHU, etc.

  • 05/14/2020: give a talk at 2020 CVPR paper sharing workshop held by MSRA [link]

  • 03/06/2020: 3 ICME papers accepted

  • 03/04/2020: 1 INFOCOM demo paper accepted

  • 02/27/2020: Join the Excutive Area Chair Committee for VALSE

  • 02/25/2020: Join the ACM Multimedia 2020 Reproducibility Committee

  • 02/24/2020: 1 CVPR paper accepted

  • 02/18/2020: 1 TSP paper accepted

  • 02/10/2020: 1 ICASSP paper accepted

  • 01/30/2020: invited as an Area Chair for ACM Multimedia 2020

  • 12/09/2019: give a talk at JD on Graph Neural Networks

  • 12/06/2019: invite Prof. Zhu Li for a talk

  • 10/21/2019: special session proposal for ICME 2020 accepted

  • 08/22/2019: Invited to visit Ryerson University, Toronto

  • 08/08/2019-08/23/2019: Invited to visit York University, Toronto, Canada

  • 07/21/2019: give a talk at USTC on "Graph Convolutional Neural Networks: from perspective of graph signal processing" [Slides]

  • 07/02/2019: 2 ACM Multimedia papers accepted

  • 05/01/2019: 1 ICIP paper accepted

  • 03/10/2019: 1 ICME paper accepted

  • 03/08/2019: 1 TIP paper accepted

  • 10/22/2018: attend ACM MM at Seoul, Korea

  • 10/07/2018: attend ICIP at Athens, Greece

  • 09/07/2018: Prof. Chia-Wen Lin visited us and gave a talk

  • 09/06/2018: 1 AAAI paper submitted

  • 09/04/2018: 1 TIP paper submitted

  • 07/02/2018: 1 ACM MM paper accepted

  • 06/29/2018: 1 GlobalSIP paper submitted

  • 06/22/2018: Xiang Gao won Wangxuan Scholarship, Peking University

  • 05/14/2018 - 05/18/2018: Invited to visit National Institute of Informatics, Tokyo

  • 05/04/2018: 2 ICIP'18 papers accepted

  • 04/27/2018: 1 GSP workshop abstract accepted

  • 04/15/2018: 1 BigMM paper submitted

  • 04/08/2018: 1 ACM MM paper submitted

  • 03/31/2018: 1 GSP workshop abstract submitted

  • 03/19/2018: 1 ACM TOMM paper accepted

  • 03/08/2018: 1 NSFC funding proposal submitted

  • 02/07/2018: 2 ICIP'18 papers submitted

  • 01/30/2018: 1 ICASSP'18 paper accepted

  • 12/15/2017: 1 ICME'18 paper submitted

  • 11/25/2017: MSRA collaborative research proposal accepted

  • 11/06/2017 - 11/10/2017: Invited to visit National Institute of Informatics, Tokyo

  • 10/27/2017: 1 ICASSP'18 paper submitted

  • 10/11/2017: 1 Alibaba Innovative Research (AIR) proposal accepted

  • 09/30/2017: 1 MSRA collaborative research proposal submitted

  • 09/01/2017: One Ph.D. student Xiang Gao and one Master student Zeqing Fu joined our group

  • 08/23/2017: 1 ACM TOMM paper submitted

  • 06/21/2017: 1 MMSP'17 paper accepted

  • 06/15/2017: Selected to be Ph.D. supervisor

  • 04/24/2017: Join Peking University as Assistant Professor

My research interests include Graph Signal Processing (GSP), Graph Neural Network (GNN), as well as the intersection between GSP and GNN for interpretable graph-based machine learning, as described below. Please refer to our overview paper on "Graph Signal Processing for Geometric Data and Beyond: Theory and Applications" [arXiv] for more discussions.

We apply the proposed GSP/GNN paradigms to the processing, analysis and synthesis of geometric data, images, brain signals, etc., with the current focus on geometric data such as 3D point clouds. Below are some major problems we address.

1. Point cloud restoration

Point cloud restoration is an inverse problem to reconstruct point clouds from degraded versions, including denoising, inpainting, upsampling, etc.. As graphs provide structure-adaptive, accurate, and compact representations for geometric data, we focus on point cloud restoration via graph signal processing and graph-based machine learning.

Point cloud denoising Point cloud upsampling

 

Selected Relevant Papers:

  • Haolan Chen, Bi'an Du, Shitong Luo, Wei Hu, "Deep Point Set Resampling via Gradient Fields," accepted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), May 2022.
  • Wei Hu, Xiang Gao, Gene Cheung, Zongming Guo, "Feature Graph Learning for 3D Point Cloud Denoising," IEEE Transactions on Signal Processing (TSP), vol. 68, pp. 2841-2856, February 2020.
  • Wei Hu, Qianjiang Hu, Zehua Wang, Xiang Gao, "Dynamic Point Cloud Denoising via Manifold-to-Manifold Distance," IEEE Transactions on Image Processing (TIP), vol. 30, pp. 6168-6183, July, 2021.
  • Shitong Luo, Wei Hu, "Score-Based Point Cloud Denoising," International Conference on Computer Vision (ICCV), 2021.

 

2. Point cloud classification / segmentation

We focus on self-supervised/weakly-supervised/robust graph representation learning for point cloud classification and segmentation.

Point cloud classification (ModelNet) Point cloud segmentation (ShapeNet)

 

Selected Relevant Papers:

  • Xiang Gao, Wei Hu, Guo-Jun Qi, "GraphTER: Unsupervised Learning of Graph Transformation Equivariant Representations via Auto-Encoding Node-wise Transformations," IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, Washington, June 2020.
  • Xiang Gao, Wei Hu, Guo-Jun Qi, "Self-Supervised Graph Representation Learning via Topology Transformations," accepted to IEEE Transactions on Knowledge and Data Engineering (TKDE), December, 2021. 
  • Daizong Liu, Wei Hu, "Imperceptible Transfer Attack and Defense on 3D Point Cloud Classification," accepted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), July 2022. 
  • Gusi Te, Wei Hu, Amin Zheng, Zongming Guo, "RGCNN: Regularized Graph CNN for Point Cloud Segmentation," ACM International Conference on Multimedia (ACM MM), Seoul, Republic of Korea, October 2018.

 

3. Point cloud generation

Learning generative models for point clouds is powerful in unsupervised representation learning to characterize the data distribution, which lays the foundation for various tasks such as shape completion, upsampling, synthesis, etc.

 

Selected Relevant papers:

  • Shitong Luo, Wei Hu, "Diffusion Probabilistic Models for 3D Point Cloud Generation," IEEE Conference on Computer Vision and Pattern Recognition (CVPR), June 2021. [Best Paper Candidate, 1 of 32 chosen from 7015 submissions]

 

4. Point cloud compression

The large amount of data in 3D point clouds significantly increase the burden for transmission and storage, especially with multiple attributes on each point. Hence, it is quite challenging to represent point clouds compactly, and efficient point cloud compression is required.

 

Selected Relevant papers:

  • Yiqun Xu, Wei Hu, Shanshe Wang, Xinfeng Zhang, Shiqi Wang, Siwei Ma, Zongming Guo, Wen Gao, "Predictive Generalized Graph Fourier Transform for Attribute Compression of Dynamic Point Clouds," IEEE Transactions on Circuits and Systems for Video Technology (TCSVT), vol. 31, no. 5, pp. 1968-1982, May 2021.

 

Current Ph.D. / Master Students

Daizong Liu Zehua Wang Qianjiang Hu Haolan Chen
Bi'an Du  Zhimin Zhang Pufan Li   Wencan Huang
   
Junyi Yao Tianshu Shen    

Co-supervised Ph.D. Student

     
Yang Liu      

Alumni

   
Xiang Gao Gusi Te    


 

  • Introduction to Computation (A), Teaching Assistant, Fall Semester 2017
  • Graph Signal Processing, Graduate Course, Fall Semester 2018
  • Graph Neural Networks, Graduate Course, Fall Semester 2019
  • Graph Signal Processing, Graduate Course, Fall Semester 2019
  • Graph Neural Networks, Graduate Course, Fall Semester 2020
  • Graph Signal Processing, Graduate Course, Fall Semester 2020
  • Graph Neural Networks, Graduate Course, Fall Semester 2021
  • Graph Signal Processing, Graduate Course, Fall Semester 2021
  • Data Structure and Algorithms, Spring Semester 2022
  • Graph Neural Networks, Graduate Course, Fall Semester 2022
  • Graph Signal Processing, Graduate Course, Fall Semester 2022

Associate Editor for the following Journals

  • IEEE Signal Processing Magazine
  • IEEE Transactions on Signal and Information Processing over Networks
  • Frontiers in Signal Processing

Members of the following Technical Committees

  • IEEE Multimedia Systems & Applications Technical Committee (MSA-TC) Member
  • IEEE Multimedia Signal Processing Technical Committee (MMSP-TC) Member

Chairs for the following Conferences

  • Tutorial Co-Chair for International Conference on Visual Communications and Image Processing (VCIP) 2022
  • Co-organizer of ICCV 2021 Workshop on "When Graph Signal Processing meets Computer Vision" [link] [PDF]
  • Special Session Co-organizer for International Conference on Image Processing (ICIP) 2021 [link]
  • Open Source Co-Chair for International Conference on Multimedia and Expo (ICME) 2021
  • Area Chair for ACM International Conference on Multimedia (ACM MM) 2020
  • Special Session Co-organizer for International Conference on Multimedia and Expo (ICME) 2020 [link]
  • Area Chair for International Conference on Multimedia and Expo (ICME) 2020
  • Excutive Area Chair for VALSE since 2020
  • Publicity Chair for IEEE International Conference on Multimedia Big Data 2020
  • Session Chair for International Conference on Multimedia and Expo (ICME) 2019
  • Session Chair for IEEE International Conference on Image Processing (ICIP) 2019

Reviewer for the following Journals/Conferences

  • IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
  • IEEE Transactions on Image Processing (TIP)
  • IEEE Transactions on Signal Processing (TSP)
  • IEEE Transactions on Multimedia (TMM)
  • IEEE Transactions on Circuits and Systems for Video Technology (TCSVT)
  • IEEE Signal Processing Letters (SPL)
  • IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
  • ACM International Conference on Multimedia (ACM MM)
  • AAAI Conference on Artificial Intelligence (AAAI), PC member
  • IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
  • IEEE International Conference on Image Processing (ICIP)
  • International Conference on Multimedia and Expo (ICME)
  • International Workshop on Multimedia Signal Processing (MMSP)