I am a Professor at the University of Science and Technology of China, where I lead the Spatial Intelligence Lab (SPIN). I am also a Guest Senior Researcher in the Computer Vision Group at the Technical University of Munich, working with Prof. Daniel Cremers.
I obtained my PhD from TUM, supervised by Prof. Uwe Stilla and Prof. Daniel Cremers. I was also a visiting PhD student in the Visual Geometry Group at the University of Oxford, working with Dr. João Henriques.
Join SPIN! We are an international research team working closely with the Technical University of Munich, University of Cambridge, and University of Oxford. Students can engage in internationally co-mentored research, including the TUM–Cambridge project Semantic Superquadric Splatting for Structured 3D Scenes. We welcome motivated Bachelor’s, Master’s, and PhD students interested in 3D Vision, Localization and Navigation, Autonomous Driving, and Embodied AI. To explore opportunities with us, please email me with your CV, transcript, and a brief description of your research interests.
SemCityLoc has been accepted to ECCV 2026!
We are organizing the TwinWorld Workshop at ECCV 2026!
Text2Loc++ has been accepted to TPAMI 2026!
C-LaV has been accepted to CVPR 2026!
TARGO has been accepted to IJCV 2026!
We organized the Multi-modal Localization and Mapping (MuLMa) Workshop at ICCV 2025!
TrafficLoc has been accepted to ICCV 2025!
SparseAlign has been accepted to CVPR 2025!
Localizing Events in Videos with Multimodal Queries has been accepted to CVPR 2025!
MonoCT has been accepted to ICRA 2025!
We aim to develop spatial intelligence that can localize, perceive, and reason in the 3D world—connecting multimodal sensing, geometry, and learning for robotics and autonomous systems.
† corresponding author · * equal contribution
IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI), 2026
Generalizing natural-language localization across diverse large-scale 3D point-cloud environments.
International Journal of Computer Vision (IJCV), 2026
A benchmark and occlusion-aware model for target-driven robotic grasping in cluttered scenes.
Conference on Computer Vision and Pattern Recognition (CVPR), 2026
Weather-robust LiDAR place recognition through conditional latent velocity-field denoising.
International Conference on Computer Vision (ICCV), 2025
Localizing traffic surveillance cameras within 3D reference maps using geometry-guided cross-modal matching.
Conference on Computer Vision and Pattern Recognition (CVPR), 2025
A fully sparse framework for accurate and efficient cooperative 3D object detection.
Conference on Computer Vision and Pattern Recognition (CVPR), 2025
Localizing video events through flexible multimodal queries that combine language and visual cues.
Conference on Computer Vision and Pattern Recognition (CVPR), 2024
Localizing a natural-language description directly within a large-scale 3D point cloud.
International Conference on Computer Vision (ICCV), 2023
Cross-attention features for robust place recognition from a single 3D scan.
Conference on Computer Vision and Pattern Recognition (CVPR Oral), 2021
Learning orientation-aware point-cloud descriptors with self-attention for place recognition.
