Angjoo Kanazawa
Berkeley EECS associate professor whose lab builds the open-source nerfstudio and gsplat libraries
- Associate Professor, Department of Electrical Engineering and Computer Sciences; leads the KAIR lab, University of California, Berkeley1
- Amazon Scholar, Frontier AI & Robotics team, Amazon1
Angjoo Kanazawa is an Associate Professor in the Department of Electrical Engineering and Computer Sciences at the University of California, Berkeley, and leads the Kanazawa AI Research (KAIR) lab under BAIR. Angjoo Kanazawa is a co-author of the papers introducing nerfstudio, a modular framework for neural radiance field development, and gsplat, an open-source library for Gaussian splatting; both projects are released under the Apache-2.0 license. Angjoo Kanazawa is also an Amazon Scholar on the Frontier AI & Robotics team, and was previously a Research Scientist at Google Research.12345
Career
Kanazawa completed a PhD in computer science at the University of Maryland, College Park, advised by David Jacobs, and was then a BAIR postdoctoral researcher at UC Berkeley advised by Jitendra Malik, Alexei A. Efros and Trevor Darrell. Kanazawa also worked as a Research Scientist at Google Research before leading the Kanazawa AI Research (KAIR) lab at Berkeley. Kanazawa's homepage notes advising companies such as Luma AI and Wonder Dynamics and a current role as an Amazon Scholar.1
Research
Kanazawa's research sits between computer vision, computer graphics and machine learning, with the goal of capturing and understanding the dynamic 3D world from everyday photos and video. As first author of "End-to-end Recovery of Human Shape and Pose" (2017), Kanazawa introduced Human Mesh Recovery (HMR), which reconstructs a full 3D body mesh from a single RGB image. The 2024 paper "Shape of Motion", which Kanazawa co-authored, reconstructs generic dynamic scenes from casually captured monocular video.167
With students and collaborators, Kanazawa has worked on faster representations for radiance fields. PlenOctrees (2021) render neural radiance fields in real time using an octree-based representation, and Plenoxels (2021) represent a scene as a sparse 3D grid with spherical harmonics, optimized without any neural network components.89
Open-source work
Kanazawa's group maintains open-source tools for 3D capture and rendering. Nerfstudio, described in a 2023 paper, is a modular framework for developing neural radiance field methods, and gsplat, described in 2024, is a library for Gaussian splatting; both are released under the Apache-2.0 license. Kanazawa's homepage lists nerfstudio, gsplat and viser among the group's open-source tools.24351
Work
nerfstudio (external site: github.com)Open-source code
Co-author of the nerfstudio paper (2023), which presents the Apache-2.0 framework for neural radiance field development; Angjoo Kanazawa's homepage describes it as one of the lab's open-source tools.241
gsplat (external site: github.com)Open-source code
Co-author of the gsplat paper (2024), which presents the Apache-2.0 library for Gaussian splatting; the homepage lists gsplat among the lab's open-source tools.351
Career timeline
Dec 2017
Posted "End-to-end Recovery of Human Shape and Pose", introducing Human Mesh Recovery, as first author.6
Mar 2021
Co-authored PlenOctrees, a method for real-time rendering of neural radiance fields.8
Dec 2021
Co-authored Plenoxels, a radiance field representation that uses no neural networks.9
Feb 2023
Co-authored the paper introducing nerfstudio, a modular framework for neural radiance fields.2
Jul 2024
Co-authored "Shape of Motion", a method for 4D reconstruction from a single video.7
Sep 2024
Co-authored the paper describing gsplat, an open-source library for Gaussian splatting.3
Other documented roles
Roles the sources document but that are not confirmed as current.
- Research Scientist, Google Research1
Honors and recognition
Related people
More sourced profiles at the same organization, in the same field, or from the same school.
- Abdul Fatir AnsariAlso at Amazon Web Services
- Aditya RameshImage, video, and 3D generation
- Adam GleaveAlso from UC Berkeley
- Alexei A. EfrosAlso at University of California, Berkeley
- Alex NicholImage, video, and 3D generation
- Adina WilliamsAlso from NYU
- Ali GhodsiAlso at University of California, Berkeley
- Amir HertzAlso at Google
- Alex SmolaAlso at Amazon Web Services
- Anca DraganAlso at University of California, Berkeley
- Ben MildenhallAlso at Google Research
- Amanda AskellAlso from NYU
Is this you?
Profiles cover public professional work only, and every line cites a source. If something is out of date, missing or wrong, send a correction with a link to an official page and we will update it. Suggest a correction
Add this badge to your website or GitHub profile README. It links back to this page. It is not an endorsement or a rating.
Markdown (for a README)
[](https://unitedstatesofamericasuperintelligence.com/people/angjoo-kanazawa/)
HTML
<a href="https://unitedstatesofamericasuperintelligence.com/people/angjoo-kanazawa/"><img src="https://unitedstatesofamericasuperintelligence.com/badges/profile.svg" alt="Angjoo Kanazawa — Profile on USASI" height="20"></a>
This profile covers public professional work only, from the sources below. It is not an endorsement or a ranking. To correct or remove it, use the correction form.