Jiahui Lei 雷嘉晖

I'm a CS Ph.D Student (2020-2025 expected) at University of Pennsylvania. My advisor is Prof. Kostas Daniilidis.

My PhD research focuses on the representation, reconstruction, analysis, and generation of dynamic 3D scenes and objects (4D), with the long-term goal of replicating and simulating the dynamic physical real world -- the Generative 4D World Models.
Many of my works embed geometric inductive biases ("structures" and "symmetries") into modeling and modern data-driven systems.
I pursue inspiring, useful, and elegant research.

I'm currently a student researcher at Google DeepMind with Prof. Leo Guibas.

I received my bachelor's degree (2016-2020) in Automation with ranking 1st/141 in the class of 2020 and with honors from Chu Kochen Honors College Mixed-Class, Zhejiang University.

New: I plan to graduate in 2025 and am seeking full-time, publication-oriented research positions. Please feel free to reach out!

Email: leijh [AT] cis [dot] upenn [dot] edu / Google Scholar / Github

Selected Research

(first, co-first, and corresponding)

	MoSca: Dynamic Gaussian Fusion from Casual Videos via 4D Motion Scaffolds Jiahui Lei, Yijia Weng, Adam W. Harley, Leonidas Guibas, Kostas Daniilidis Arxiv, 2024 project page / paper (NEW!) / arXiv / video (NEW!) / code (Available Now!)
	DIMO: Diverse 3D Motion Generation for Arbitrary Objects Linzhan Mou, Jiahui Lei^†, Chen Wang, Lingjie Liu, Kostas Daniilidis († corresonding author) Under review, coming soon paper / video
	HDGS: Textured 2D Gaussian Splatting for Enhanced Scene Rendering Yunzhou Song, Heguang Lin, Jiahui Lei^†, Lingjie Liu, Kostas Daniilidis (* equal contribution; † corresonding author) Arxiv, 2024 project page / paper / video
	GART: Gaussian Articulated Template Models Jiahui Lei, Yufu Wang, Georgios Pavlakos, Lingjie Liu, Kostas Daniilidis CVPR, 2024 project page / arXiv / video (YouTube) / video (Bilibili) / code
	Track Everything Everywhere Fast and Robustly Yunzhou Song, Jiahui Lei, Ziyun Wang, Lingjie Liu, Kostas Daniilidis (* equal contribution) ECCV, 2024 project page / arXiv / video (YouTube) / code
	DynMF: Neural Motion Factorization for Real-time Dynamic View Synthesis with 3D Gaussian Splatting Agelos Kratimenos, Jiahui Lei^†, Kostas Daniilidis († corresonding author) ECCV, 2024 project page / arXiv
	NAP: Neural 3D Articulation Prior Jiahui Lei, Congyue Deng, Bokui Shen, Leonidas Guibas, Kostas Daniilidis NeurIPS, 2023 project page / arXiv / video (YouTube) / code
	Banana: Banach Fixed-Point Network for Pointcloud Segmentation with Inter-Part Equivariance Congyue Deng, Jiahui Lei, Bokui Shen, Kostas Daniilidis, Leonidas Guibas (* equal contribution) NeurIPS, 2023 project page / arXiv / video (YouTube)
	EFEM: Equivariant Neural Field Expectation Maximization for 3D Object Segmentation Without Scene Supervision Jiahui Lei, Congyue Deng, Karl Schmeckpeper, Leonidas Guibas, Kostas Daniilidis CVPR, 2023 project page / dataset / arXiv / video (YouTube) / video (Bilibili) / code
	CaDeX: Learning Canonical Deformation Coordinate Space for Dynamic Surface Representation via Neural Homeomorphism Jiahui Lei, Kostas Daniilidis CVPR, 2022 project page / arXiv / video (YouTube) / video (Bilibili) / code
	Unified Fourier-based Kernel and Nonlinearity Design for Equivariant Networks on Homogeneous Spaces Yinshuang Xu, Jiahui Lei, Edgar Dobriban, Kostas Daniilidis (* equal contribution) ICML, 2022 paper
	Pix2Surf: Learning Parametric 3D Surface Models of Objects from Images Jiahui Lei, Srinath Sridhar, Paul Guerrero, Minhyuk Sung, Niloy Mitra, Leonidas Guibas ECCV, 2020 project page / arXiv / video / poster / bibtex / code