Shihang Wei 魏诗航

PhD Student, Purdue University · advised by Fengqing Maggie Zhu

I build fast, compact 4D reconstructions of the dynamic world.

My research interests are in 3D and 4D reconstruction, with a focus on building solid, high-fidelity dynamic 3D/4D models efficiently. I am especially interested in approaches that can reduce reconstruction time while preserving quality and reducing the model storage size, whether through feed-forward methods, per-scene optimization, or hybrid systems. I also love photography and hiking, check out my Instagram to see my latest work.

Publications

• DanceNet3D: A 3D Dance Dataset with Multi-View Videos and 3DGS Reconstructions. Shihang Wei*, Mingjian Li*, Ran Gong, Yueyu Hu, Yao Wang (* equal contribution). CVPRW 2026. Best Paper Award

News

• Aug 2026 — Joined Purdue's ECE PhD program, working with Professor Fengqing Maggie Zhu.
• Jun 2026 — Received the Best Paper Award at the 3DMV Workshop at CVPR 2026 for DanceNet3D.
• Jun 2026 — Presented our DanceNet3D dataset poster at the 3DMV workshop, CVPR 2026.
• May 2026 — Received my BS in Computer Science from NYU Tandon SOE.
• Apr 2026 — Our dataset DanceNet3D was published — huge thanks to my research advisor Yao Wang for her guidance throughout the project.

Contact

shihang.wei@purdue.edu · West Lafayette, IN

https://scholar.google.com/citations?view_op=list_works&hl=en&user=6sqBIDkAAAAJ · https://orcid.org/0009-0006-3034-2230 · https://github.com/shihangvicwei · https://www.linkedin.com/in/shihangweivictor · https://www.instagram.com/vicccww/