|
Guangzhao (Alex) He I'm a first-year CS Ph.D. student at Cornell University, working with Prof. Wei-Chiu Ma and Prof. Hadar Elor. I'm also a student researcher at Google DeepMind, hosted by Dor Verbin. I received my Bachelor's degree with honors in Computer Science at Chu Kochen Honors College, Zhejiang University. I've had the honor of working closely with Prof. Xiaowei Zhou and Prof. Sida Peng at Zhejiang University, and spent a lovely summer at Stanford as a research intern working with Prof. Jiajun Wu and Prof. Elliott Wu. I produce music under both aliases AlexHe and Elykz, experimenting among the genres of House, Hip-hop and New-age. My original music has gathered over 2 million streams across platforms. Email / CV / Google Scholar / X / Github 
|
|
News
|
PublicationMy research interests lie at the intersection of Computer Vision and Reasoning, particularly in the representation, understanding and long-term memory of dynamic 3D scenes, with the goal of enabling machines to perceive the dynamic and diverse world as humans do. |
|
Category-Agnostic Neural Object Rigging
Guangzhao He*, Chen Geng*, Shangzhe Wu, Jiajun Wu CVPR 2025 project page / arXiv A feed-forward representation that encodes deformable 4D objects into a sparse set of spatially-grounded blobs to disentangle pose and instance information, enabling intuitive pose manipulation while preserving rich instance-specific information. |
|
ERNet: Efficient Non-Rigid Registration Network for Point Sequences
Guangzhao He, Yuxi Xiao, Sida Peng, Zhen Xu, Xiaowei Zhou ICCV 2025 project page / arXiv A feed-forward pipeline that registers any object template onto a sequence of non-rigidly deforming point clouds, with 4.6x faster inference speed and 2x better accuracy than previous SOTA. |
|
4K4D: Real-Time 4D View Synthesis at 4K Resolution
Zhen Xu, Sida Peng, Haotong Lin, Guangzhao He, Jiaming Sun, Yujun Shen, Hujun Bao, Xiaowei Zhou CVPR 2024 project page / arXiv A novel point-based 4D representation reconstructed with multi-view videos, which can be rendered at 4k 200+ FPS with state-of-the-art quality on a single RTX 4090. |
|
EasyVolcap: Accelerating Neural Volumetric Video Research
Zhen Xu, Tao Xie, Sida Peng, Haotong Lin, Qing Shuai, Zhiyuan Yu, Guangzhao He, Jiaming Sun, Hujun Bao, Xiaowei Zhou SIGGRAPH Asia 2023 TC project page / arXiv A Python & PyTorch library for accelerating volumetric video research, particularly in the area of neural dynamic scene representation, reconstruction, and rendering. |
Academic ServiceConference Reviewer: ECCV (2026), CVPR (2026), ICLR (2026), ICCV (2025), 3DV (2026). |
|
Template by Jon Barron Last updated: |