Selected Projects
SAFE-EQA: Semantic-Aware Efficient Exploration for Embodied Question Answering
Kai Xu et al.
ECCV 2026
“Guide embodied agents toward question-relevant regions by combining semantic maps, efficient exploration, and VLM reasoning.”
HoloTetSphere: Unified TetSphere Mesh Reconstruction for Physical Simulations
YaQiao Dai, Renjiao Yi, Zhirui Gao, Wei Chen, Kai Xu, Chenyang Zhu
ECCV 2026
“Topology-adaptive reconstruction of unified, single-connected tetrahedral meshes for reliable downstream physical simulation.”
RoboBPP: Benchmarking Robotic Online Bin Packing with Physics-based Simulation
Zhoufeng Wang, Hang Zhao, Juzhan Xu, Shishun Zhang, Zeyu Xiong, Ruizhen Hu, Chenyang Zhu, Zecui Zeng, Kai Xu
Submitted 2026
“A physics-based benchmark that finally asks whether a packing plan is physically feasible.”
From Specialist to Generalist: A Comprehensive Survey on World Models
Kai Xu, Hang Zhao, Ruizhen Hu, Yuhang Huang, Ziqiao Zhou, Wancheng Feng, Li Yi, Sida Peng, et al.
Computational Visual Media 2026
“Specialist and generalist world models must coexist — precision and generalization pull in opposite directions.”
Designing Pin-pression Gripper and Learning its Dexterous Grasping with Online In-hand Adjustment
Hewen Xiao, Xiuping Liu, Hang Zhao, Jian Liu, Kai Xu
SIGGRAPH 2025 · ACM TOG
“A gripper inspired by pin-pression toys: each finger reshapes itself around the object.”
RemixFusion: Residual-based Mixed Representation for Large-scale Online RGB-D Reconstruction
Yuqing Lan, Chenyang Zhu, Shuaifeng Zhi, Jiazhao Zhang, Zhoufeng Wang, Renjiao Yi, Yijie Wang, Kai Xu
SIGGRAPH Asia 2025 · ACM TOG
“High-quality, large-scale online RGB-D reconstruction from a residual-based mixed representation.”
MIPS-Fusion: Multi-Implicit-Submaps for Scalable and Robust Online Neural RGB-D Reconstruction
Yijie Tang, Jiazhao Zhang, Zhinan Yu, He Wang, Kai Xu
SIGGRAPH Asia 2023 · ACM TOG
“Divide-and-conquer neural SLAM: many small implicit submaps instead of one big field.”
ROSEFusion: Random Optimization for Online Dense Reconstruction under Fast Camera Motion
Jiazhao Zhang, Chenyang Zhu, Lintao Zheng, Kai Xu
SIGGRAPH 2021 · ACM TOG
“Real-time RGB-D reconstruction that survives violently fast camera motion.”
SymmetryNet: Learning to Predict Reflectional and Rotational Symmetries of 3D Shapes from Single-View RGB-D Images
Yifei Shi, Junwen Huang, Hongjia Zhang, Xin Xu, Szymon Rusinkiewicz, Kai Xu
SIGGRAPH Asia 2020 · ACM TOG
“End-to-end learnable symmetry prediction from a single RGB-D view.”
Multi-Robot Collaborative Dense Scene Reconstruction
Siyan Dong, Kai Xu, Qiang Zhou, Andrea Tagliasacchi, Shiqing Xin, Matthias Nießner, Baoquan Chen
SIGGRAPH 2019 · ACM TOG
“A fleet of robots that plans, divides and conquers an unknown indoor scene together.”
SCORES: Shape Composition with Recursive Substructure Priors
Chenyang Zhu, Kai Xu, Siddhartha Chaudhuri, Renjiao Yi, Hao Zhang
SIGGRAPH Asia 2018 · ACM TOG
“Composing a coherent shape out of incompatible parts, recursively.”
Learning to Group and Label Fine-Grained Shape Components
Xiaogang Wang, Bin Zhou, Haiyue Fang, Xiaowu Chen, Qinping Zhao, Kai Xu
SIGGRAPH Asia 2018 · ACM TOG
“Stock 3D models already encode the artist's decomposition — don't throw those components away.”
Semi-Supervised Co-Analysis of 3D Shape Styles from Projected Lines
Fenggen Yu, Yan Zhang, Kai Xu, Ali Mahdavi-Amiri, Hao Zhang
SIGGRAPH 2018 · Graphics Replicability Stamp
“Localizing style patches on 3D shapes with only weak supervision.”
Autonomous Reconstruction of Unknown Indoor Scenes Guided by Time-varying Tensor Fields
Kai Xu, Lintao Zheng, Zihao Yan, Guohang Yan, Eugene Zhang, Matthias Nießner, Oliver Deussen, Daniel Cohen-Or, Hui Huang
SIGGRAPH Asia 2017 · ACM TOG
“Tensor fields as a global planner for a robot exploring an unknown room.”
Deformation-Driven Shape Correspondence
Chenyang Zhu, Renjiao Yi, Wallace Lira, Ibraheem Alhashim, Kai Xu, Hao Zhang
SIGGRAPH 2017 · ACM TOG
“Structure-aware correspondence found by deforming one shape into the other.”
Autoscanning for Coupled Scene Reconstruction and Proactive Object Analysis
Kai Xu, Hui Huang, Yifei Shi, Hao Li, Pinxin Long, Jianong Caichen, Wei Sun, Baoquan Chen
SIGGRAPH Asia 2015 · ACM TOG
“A robot that reconstructs a scene and pokes at its objects to understand them.”
Deformation-Driven Topology-Varying 3D Shape Correspondence
SIGGRAPH Asia 2015 · ACM TOG
“Corresponding two shapes whose part topologies do not even match.”
Organizing Heterogeneous Scene Collections through Contextual Focal Points
Kai Xu, Rui Ma, Hao Zhang, Chenyang Zhu, Ariel Shamir, Daniel Cohen-Or, Hui Huang
SIGGRAPH 2014 · ACM TOG
“Finding the recurring ‘focal points’ that organize a messy collection of 3D scenes.”
Topology-Varying 3D Shape Creation via Structural Blending
SIGGRAPH 2014 · ACM TOG
“Blending two shapes through their structures, letting topology change along the way.”
Layered Analysis of Irregular Facades via Symmetry Maximization
SIGGRAPH 2013 · ACM TOG
“Peeling an irregular building facade into layers by maximizing symmetry.”
Co-Hierarchical Analysis of Shape Structures
SIGGRAPH 2013 · ACM TOG
“One consistent hierarchy shared across a whole set of structurally different shapes.”
Multi-Scale Partial Intrinsic Symmetry Detection
SIGGRAPH Asia 2012 · ACM TOG
“Symmetry is scale-dependent — so detect it across scales, and only partially.”
Photo-Inspired Model-Driven 3D Object Modeling
Kai Xu, Hanlin Zheng, Hao Zhang, Daniel Cohen-Or, Ligang Liu, Yueshan Xiong
SIGGRAPH 2011 · ACM TOG
“Turning a single photograph into a 3D model by deforming a structured template.”
Style-Content Separation by Anisotropic Part Scales
SIGGRAPH Asia 2010 · ACM TOG
“Style lives in the anisotropic scales of parts; content lives in what the parts are.”
Partial Intrinsic Reflectional Symmetry of 3D Shapes
Kai Xu, Hao Zhang, Andrea Tagliasacchi, Ligang Liu, Guo Li, Min Meng, Yueshan Xiong
SIGGRAPH Asia 2009 · ACM TOG
“A voting scheme that turns partial symmetries into an intrinsic reflectional symmetry axis transform.”
The complete publication list, including papers not shown here, is on the publications page.