Weiyi Xue’s Homepage

💫About Me

I am currently a second-year M.S.E. student at GEAI Lab of Tongji University and advised by Prof. Guang Chen. Before this, I obtained the bachelor’s degree in Vehicle Engineering at Tongji University.

My research background is primarily in 3D computer vision, with core experience in point cloud registration, 3D scene reconstruction and generation. In the short term, my research centers on investigating the intersection of generative models and 3D vision, with a specific focus on tasks such as 3D generation and scene editing. Looking further ahead, I am deeply interested in the potential of world models and their applications, such as in simulation and the integration with vision-language-action (VLA) systems.  

🔥News

[2025/09] Invited as a Reviewer for ICLR, CVPR.
[2025/09] 🎉 One paper accepted to NeurIPS 2025.
[2025/07] Invited as a Reviewer for AAAI.
[2025/05] Invited as a Reviewer for NeurIPS.
[2025/05] Started an internship at DJI, machine learning group, Shenzhen, China.
[2024/09] 🎉 One paper on pose-free reconstruction accepted to NeurIPS 2024.
[2024/02] 🎉 One paper on dynamic scene reconstruction accepted to CVPR 2024.
[2023/10] 🎉 One paper on point cloud registration accepted to WACV 2024.
[2023/06] Graduated with a Bachelor’s degree from Tongji University, Shanghai, China.

📝Publications

	"GeoNLF: Geometry guided Pose-Free Neural LiDAR Fields" Weiyi Xue, Zehan Zheng, Fan Lu, Haiyun Wei, Guang Chen†, Changjun Jiang. NeurIPS 2024 Arxiv Poster Code We explore the geometric insights of point clouds, which provide explicit registration priors for reconstruction. Based on this, we propose Geometry guided Neural LiDAR Fields(GeoNLF), a hybrid framework performing alternately global neural reconstruction and pure geometric pose optimization.
	"LiDAR4D: Dynamic Neural Fields for Novel Space-time View LiDAR Synthesis" Zehan Zheng, Fan Lu, Weiyi Xue, Guang Chen†, Changjun Jiang. CVPR 2024 Arxiv Poster Code Slides Differentiable LiDAR-only framework for novel space-time LiDAR view synthesis, which reconstructs dynamic driving scenarios and generates realistic LiDAR point clouds end-to-end. It also supports simulation in the dynamic scene.
	"HDMNet: AHierarchical Matching Network with Double Attention for Large-scale Outdoor LiDAR Point Cloud Registration" Weiyi Xue, Fan Lu, Guang Chen†. WACV 2024 Paper We proposed a feature consistency enhanced double-soft matching network for large-scale outdoor scene point cloud registration, achieving two-stage matching with high flexibility while enlarging the receptive field in a patch-to-patch manner.

💻Research Experience

• July 2023 - Present
  Research Assistant - Generalist Embodied AI Lab (GEAI Lab), Tongji University
  Advisor: Prof. Guang Chen
  Research included: 3D/4D reconstruction, Diffusion Models

• April 2025 - Present
  Research Intern - Machine Learning Group, DJI
  Research included: MLLMs, 3D/4DGS

• July 2023 - Present
  Research Assistant - Integrated Positioning Lab (IPLab), Tongji University
  Advisor: Prof. Lu Xiong
  Research included: SLAM, Point Cloud

🛠️Engineering Experience

• 2019 - 2020
  National-level Undergraduate Innovation Training Program
  Team leader
  Self-adjusting Rearview Mirror Based on Human Eye Key Point Recognition, and successfully finalized and concluded the project. The Bronze Award in the ”Challenge Cup” National Undergraduate Extracurricular Academic Science and Technology Works Competition.

🏆Honors and Awards

• Scholar award of NeurIPS, 2025
• Outstanding Student of Tongji University, 2021
• First Prize of Tongji University Scholarship (Top 5%), 2021
• Silver Award in the China International College Students’ Innovation Competition, 2020
• The Bronze Award in the ”Challenge Cup” National Undergraduate Extracurricular Academic Science and Technology Works Competition, 2020