Wei Hu
长聘副教授
Researches 3D vision, graph machine learning, graph signal processing and AI for science.
Wei Hu's lab at Peking works on 3D vision & neural rendering, graph learning and signal processing & information theory. 43 papers since 2023. 1 PhD student. 5 alumni. Applications close .
At a glance
- papers, 2023–2026
- 11/ year
- at T1 venues
- 0%
- PhD students
- 1
- MS students
- 0
- alumni listed
- 5
Papers by yearfrom the lab homepage, 2023–2026
- 8 papers2026 · T1 0 · collecting
- 15 papers2025 · T1 0
- 7 papers2024 · T1 0
- 13 papers2023 · T1 0
T1 venuesOther
Research focuspaper titles that use each topic's words
- Graph learning14 papers
- graph8 papers
- 3D vision & neural rendering5 papers
- Signal processing & information theory3 papers
- graph signal1 paper
Paper list
From the lab homepage.
20268 papers
- accepted to IEEE Transactions on Information Forensics and Security (TIFS), .Are Large Vision-Language Models Robust to Adversarial Visual Transformations？
- accepted to IEEE International Conference on Multimedia and Expo (ICME), .From Part to Whole: 3D Generative World Model with an Adaptive Structural Hierarchy
- accepted to Transactions on Dependable and Secure Computing (TDSC), .Hard-Label Black-Box Attacks on 3D Point Clouds
- accepted to IEEE Transactions on Multimedia (TMM), .HierGF: Hierarchical Gaussian Fields via Geometry–perception Message Passing for Sparse-view 3D Reconstruction
- accepted to IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP), .Importance of Balance: Lightweight Transformer via Signed Graph Algorithm Unrolling for Eeg Signal Denoising
- accepted to the International Conference on Learning Representations (ICLR), .Lightweight Transformer for EEG Classification via Balanced Signed Graph Algorithm Unrolling
- accepted to AAAI Conference on Artificial Intelligence (AAAI)Spatial-Spectral Homogeneous Attacks on Physical-World Large Vision-Language Models
- accepted to IEEE International Workshop on Multimedia Signal Processing (MMSP), .UTID: Time-Independent Cube-Absorb Diffusion for Unified Large Scene Generation
202515 papers
- accepted to IEEE Transactions on Neural Networks and Learning Systems (TNNLS), July, .A Survey of Attacks on Large Vision-Language Models: Resources, Advances, and Future Trends
- accepted to IEEE Transactions on Neural Networks and Learning Systems, June, .A Survey on Text-guided 3D Visual Grounding: Elements, Recent Advances, and Future Directions
- accepted to IEEE Transactions on Multimedia (TMM), .Active Cross-Task Knowledge Transfer Learning Framework for Semi-supervised Joint 3D Grounding and Captioning
- accepted to ACM International Conference on Multimedia (ACM MM), .Behave Your Motion: Habit-preserved Cross-category Animal Motion Transfer
- in IEEE International Conference on Multimedia & Expo (ICME), .Denoising Diffusion Probabilistic Model for Point Cloud Compression at Low Bit-Rates
- accepted to Transactions on Multimedia Computing Communications and Applications (TOMM), February, .Dynamic Point Cloud Denoising via Gradient Fields
- accepted to ACM International Conference on Multimedia (ACM MM), .Fast3D: Accelerating 3D Multi-modal Large Language Models for Efficient 3D Scene Understanding
- accepted to International Conference on Computer Vision (ICCV), .HyperGCT: A Dynamic Hyper-GNN-Learned Geometric Constraint for 3D Registration
- in The International Conference on Artificial Neural Networks (ICANN), .Improving the Transferability of 3D Point Cloud Attack via Spectral-aware Admix and Optimization Designs
- accepted to International Conference on Computer Vision (ICCV), .Large Scene Generation with Cube-Absorb Discrete Diffusion
- accepted to International Journal of Computer Vision (IJCV), July, .Learning Latent Part-Whole Hierarchies for Point Clouds
- accepted to APSIPA Transactions on Signal and Information Processing, January, .Molecular Representation Learning via Hierarchical Graph Transformer
20247 papers
- accepted to ACM International Conference on Multimedia (ACM MM), .Advancing 3D Object Grounding Beyond a Single 3D Scene
- accepted to ACM International Conference on Multimedia (ACM MM), .Cross-Task Knowledge Transfer for Semi-supervised Joint 3D Grounding and Captioning
- accepted to IEEE Conference on Computer Vision and Pattern Recognition (CVPR), March, .Generative 3D Part Assembly via Part-Whole-Hierarchy Message Passing
- accepted to IEEE Transactions on Multimedia (TMM), July, .Imperceptible Backdoor Attacks on Text-Guided 3D Scene Grounding
- accepted to International Conference on Pattern Recognition (ICPR), .Joint Top-Down and Bottom-Up Frameworks for 3D Visual Grounding
- accepted to European Conference on Computer Vision (ECCV), September, .RangeLDM: Fast Realistic LiDAR Point Cloud Generation
- accepted to Cell Research, March, .System-level time computation and representation in the suprachiasmatic nucleus revealed by large-scale calcium imaging and machine learning
202313 papers
- accepted to International Conference on Computer Vision (ICCV), .3DHacker: Spectrum-based Decision Boundary Generation for Hard-label 3D Point Cloud Attack
- accepted to ACM International Conference on Multimedia (ACM MM)Dense Object Grounding in 3D Scenes
- accepted to IEEE Conference on Computer Vision and Pattern Recognition (CVPR), March, .Density-Insensitive Unsupervised Domain Adaption on 3D Object Detection
- accepted to the AAAI Conference on Artificial Intelligence (AAAI), December, .DHGCN: Dynamic Hop Graph Convolution Network for Self-supervised Point Cloud Learning
- accepted to the AAAI Conference on Artificial Intelligence (AAAI), December, .Explicitly Perceiving and Preserving the Local Geometric Structures for 3D Point Cloud Attack
- accepted to IEEE Transactions on Multimedia (TMM), January, .Exploring Optical-Flow-Guided Motion and Detection-Based Appearance for Temporal Sentence Grounding
- accepted to International Conference on Acoustics, Speech, and Signal Processing (ICASSP), December, .InvariantOODG: Learning Invariant Features of Point Clouds for Out-of-Distribution Generalization
- accepted to IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), November, .Point Cloud Attacks in Graph Spectral Domain: When 3D Geometry Meets Graph Signal Processing
- accepted to IEEE Transactions on Multimedia (TMM), August, .Robust Geometry-Dependent Attack for 3D Point Clouds
- accepted to IEEE International Conference on Image Processing (ICIP), .Robust Graph-based Segmentation of Noisy Point Clouds
- Asia-Pacific Signal and Information Processing Association Annual Summit and Conference (APSIPA ASC), Taipei, Taiwan, October, .Robust Watermark Imaging via Graph-Signal Optimization
- accepted to ACM Transactions on Multimedia Computing, Communications, and Applications (TOMCCAP), May, .Self-Supervised Multi-View Learning via Auto-Encoding 3D Transformations
Where alumni went5 listed
Apply to PekingFall 2027
Computer Science PhD (international applicants)
- Round
- Main
- Opens
- —
- GRE
- See program page
- English
- See program page
- Fee
- See program page
Not yet confirmed for this intake.
Similar labsnearest by research summary
- GBUCL
Francisco Porto Guerra E Vasconcelos
- 3D vision & neural rendering
- CNTsinghua
Hang Zhao
- Embodied & active vision
- CNSJTU
Weidi Xie
- Video understanding
- CNPeking
Shanghang Zhang
- Vision-language & multimodal
- USMichigan
Bernadette Bucher
- Robot perception & SLAM
- CNSJTU
Weiyao Lin
- Embodied & active vision
Your notes
Save this lab to keep notes.
Data
Some numbers could not be verified:
- The lab site and the paper record do not agree.
- Collected
- Verified
- Papers
- The lab homepage
- Members
- Lab homepage
- Alumni
- Lab homepage
Tiers are venue classifications, not lab rankings.
Pekingcloses