Weiguang Zhao (赵伟光)

Ph.D. Candidate (third year)

University of Liverpool

weiguang.zhao@liverpool.ac.uk

Github / Goolge Scholar / Youtube / Twitter

Home     Biography     Publications     Robotics     Patents     Service    

Biography

I am a third-year Ph.D. candidate in Computer Science department at University of Liverpool in UK, based off-site at Xi'an Jiaotong-Liverpool University in China. I am fortunate to be co-advised by Prof. Kaizhu Huang, Prof. Rui Zhang, Prof. Qiufeng Wang, and Prof. Guangliang Cheng. I received my Bachelor’s degree in Electronic Information Science and Technology from Heilongjiang University and my Master’s degree in MRes Pattern Recognition and Intelligent Systems from University of Liverpool. Previously, I have worked or interned at DJI, Duke (Kunshan) University and Shanghai AI Lab. Feel free to contact me via email for discussions or any potential opportunities.

Research Goal:
My long-term research goal is to develop efficient and scalable 3D/4D vision models and deployable robotic systems for embodied spatial intelligence, enabling robots to perceive, understand, and interact with the real physical world.

Research interests:
3D/4D Vision: World Model, Point Tracking, Semantic/Instance Segmentation, Reconstruction and Generation;
Embodied AI: Dexterous Hand, Robotic Arm, Drone, Humanoid Robot; DP, RL, WAM, VLA

Recent News:
[Jun. 2026] One paper has been accepted by ECCV.
[Mar. 2026] Two papers has been accepted by SIGGRAPH.
[Sep. 2025] One paper has been accepted by IJCV.
[Jul. 2025] One paper on point cloud has been accepted by ACM MM.
[Feb. 2025] Two papers on point cloud has been accepted by CVPR.
[Aug. 2024] I got the 2st place in ScanNet200 3D semantic segmentation benchmark.
[Jul. 2023] One paper on point cloud has been accepted by ICCV.
[Jan. 2023] I got the 1st place in ScanNetV2 3D instance segmentation benchmark.

Publications (Selected)

(* stands for equal contribution)

Preprints

TeleCatch

Tele-Catch: Adaptive Teleoperation for Dexterous Dynamic 3D Object Catching
W. Zhao, J. Dong, R. Zhang, K. Li, Q. Zhao, K. Huang
arXiv, 2026 ,Under Review
[ arXiv] [Code]

hand_survey

Towards Robotic Dexterous Hand Intelligence: A Survey
W. Zhao*, T. Liang*, X. Guo*, Rui Zhang, I. King, K. Huang
arXiv, 2026 ,Under Review
[ arXiv] [Project]
InfiniHand

InfiniHand: Streaming World-Space Hand Motion Estimation from Egocentric Video
K. Ren, K. Song, W. Zhao, Y. Wang, Y. Liu, B. Dai, H. Guo, C. Shen, M. Yu, T. Lu, J. Dong
arXiv, 2026 ,Under Review
[ arXiv] [ Project]
InternW0

InternW0-Δ: A World Action Model Bridging Predictive Dynamics and Actions
Physical Intelligence Team (42 Authors), Shanghai AI Laboratory
arXiv, 2026 ,Under Review
[ Technical Report] [ Project]

Conferences

SynthVerse

SynthVerse: A Large-Scale Diverse Synthetic Dataset for Point Tracking
W. Zhao*, H. Xu*, X. Miao, Q. Zhao, R. Zhang, K. Huang, N. Gao, P. Cao, M. Sun, M. Yu, T. Lu, L. Xu, J. Dong, J. Pang
ACM SIGGRAPH Annual Conference (SIGGRAPH), 2026,CCF-A, Core-A*
[ Paper] [ Project] [ Code]

TrajVG

TrajVG: 3D Trajectory-Coupled Visual Geometry Learning
X. Miao, W. Zhao, T. Lu, L. Xu, M. Yu, Y. Long, J. Pang, J. Dong
ACM SIGGRAPH Annual Conference (SIGGRAPH), 2026,CCF-A, Core-A*
[ Paper] [ Project] [ Code]

Geo-DPO

Geo-DPO: Aligning Semantic Intent with Geometry for 3D Affordance Segmentation
Z. Yang, X. Wang, X. Qiu W. Zhao, Q. Zhang, J. Xiao
European Conference on Computer Vision (ECCV), 2026,CCF-B, Core-A*
[ Paper] [Code]

3DAD

BFANet: Revisiting 3D Semantic Segmentation with Boundary Feature Analysis
W. Zhao, R. Zhang, Q. Wang, G. Cheng, K. Huang
Conference on Computer Vision and Pattern Recognition (CVPR), 2025,CCF-A, Core-A*
[ Paper] [ Code]

3DAD

PO3AD: Predicting Point Offsets toward Better 3D Point Cloud Anomaly Detection
J. Ye*, W. Zhao*, X. Yang, G. Cheng, K. Huang
Conference on Computer Vision and Pattern Recognition (CVPR), 2025,CCF-A, Core-A*
[ Paper] [ Code]

3DAD

Towards Training-Free Open-World Classification with 3D Generative Models
X. Xia*, W. Zhao*, Y. Yan, G. Yang, R. Zhang, K.Huang, X. Yang
ACM International Conference on Multimedia (ACM MM), 2025,CCF-A, Core-A*
[ Paper] [ Code]

PBNet

Divide and Conquer: 3D Point Cloud Instance Segmentation With Point-Wise Binarization
W. Zhao, Y. Yan, C. Yang, J. Ye, X. Yang, K. Huang
International Conference on Computer Vision (ICCV), 2023,CCF-A, Core-A*
[ Paper] [ Code] [ Video]

Journals

KDTalk

Unlock Pose Diversity: Accurate and Efficient Implicit Keypoint-based Spatiotemporal Diffusion for Audio-driven Talking Portrait
C. Yang, K. Yao, Y. Yan, C. Jiang, W. Zhao, J. Sun, etc.
International Journal of Computer Vision (IJCV) , 2026,CCF-A，JCR-Q1
[ Paper] [ Code] [ Demo]
Diff_OP3D

Open-Pose 3D Zero-Shot Learning: Benchmark and Challenges
W. Zhao*, G. Yang*, R. Zhang, C. Jiang, C. Yang, Y. Yan, A. Hussain, K. Huang
Neural Networks (NN), 2025,CCF-B，JCR-Q1
[ Paper] [ Code]

MPANet

Revisiting 3D point cloud analysis with Markov process
C. Jiang, W. Ma, K. Huang, Q. Wang, X. Yang,W. Zhao, J. Wu, X. Wang, etc.
Pattern Recognition (PR), 2025,CCF-B，JCR-Q1
[ Paper] [ Code]


Robotics (Selected)

Tele-Catch robotics

Tele-Catch: Adaptive Teleoperation for Dexterous Dynamic 3D Object Catching, 2025
Responsibilities: Project Holder; Coder
Keys: Teleoperation, Dexterous Hands, Point Cloud; Reinforcement Learning, Diffusion Policy

Robotic grasping

Towards Cross-device and Training-free Robotic Grasping in 3D Open World, 2022-2024
Responsibilities: Project Holder; Coder
Keys: Point Cloud; Cross-device; Training-free; Open World
DJI SDK

DJI SDK, 2018-2020
Responsibilities: Coder; Maintainer
Keys: Flight Control; Onboard SDK; Payload SDK;

VR survey vehicle

Design and Implementation of Survey Vehicle Based on VR, 2017-2018
National Undergraduate Innovation Training Program
Responsibilities: Project Holder; Coder
Keys: Gesture-based Control; Multi-machine Communication; VR Imaging;

Attitude Control and Flight of Quadcopter, 2016-2017
University Undergraduate Innovation Training Program
Responsibilities: Project Holder; Coder
Keys: Flight Control From Scratch; GPS Positioning; Ultrasonic Hover


Invention Patents

CN117456236B 3D point cloud zero-shot classification method, device, electronic device and storage medium.
CN114266860B 3D face model establishment method, device, electronic device and storage medium.
CN116580267A Generating defect samples method, device, electronic device and storage medium.
CN202510224283.6 Object removal based on 3D scenes methods, device, electronic device and medium.
CN202510224297.8 Digital human drive method, device, electronic device and storage medium.
CN202510224304.4 3D point cloud data segmentation method, device, electronic device and storage medium.


Service

Conference Review Service: CVPR, ICCV, ICLR, AAAI, ACM MM, ICRA, ICME, BMVC etc.
Journal Review Service: TPAMI, TIP, TVCG, TMM, TCSVT, TGRS, Pattern Recognition, Neural Networks, etc.

Teaching Assistants:
STATS 302: Principle of Machine Learning         Duke (Kunshan) University, 2023 Fall
ECE 590K: Introduction to Deep Learning         Duke (Kunshan) University, 2024 Spring
ECE580K: Introduction to Machine Learning         Duke (Kunshan) University, 2025 Spring