Hello, I'm Keunhong Park
I am a founding member at World Labs where I lead model pretraining. My research focuses on world models—from large-scale pretraining to real-time generation—combining diffusion models with 3D scene representations. See our recent Marble and RTFM releases.
Previously I was a research scientist at Google where I built technology to generate 3D assets for products on Google Search. I received my Ph.D from the University of Washington in 2021, advised by Ali Farhadi and Steve Seitz.
Highlights
Publications
RTFM: Real-Time Frame Model
Keunhong Park with collaborators at World Labs
Blog Post, 2025
A real-time, auto-regressive diffusion model renders persistent 3D worlds on a single GPU.
IllumiNeRF 3D Relighting without Inverse Rendering
Xiaoming Zhao, Pratul Srinivasan, Dor Verbin, Keunhong Park, Ricardo Martin-Brualla, Philipp Henzler
NeurIPS, 2024
3D relighting by distilling samples from a 2D image relighting diffusion model into a latent-variable NeRF.
ReconFusion: 3D Reconstruction with Diffusion Priors
Rundi Wu, Ben Mildenhall, Philipp Henzler, Keunhong Park, Ruiqi Gao, Daniel Watson, Pratul Srinivasan, Dor Verbin, Jonathan T. Barron, Ben Poole, Aleksander Holynski
CVPR, 2024
Using an multi-view image conditioned diffusion model to regularize a NeRF enabled few-view reconstruction.
CamP: Camera Preconditioning for Neural Radiance Fields
Keunhong Park, Philipp Henzler, Ben Mildenhall, Jonathan T. Barron, Ricardo Martin-Brualla
SIGGRAPH Asia, 2023 Journal Paper
Preconditioning camera optimization during NeRF training significantly improves their ability to jointly recover the scene and camera parameters.
HyperNeRF: A Higher-Dimensional Representation for Topologically Varying Neural Radiance Fields
Keunhong Park, Utkarsh Sinha, Peter Hedman, Jonathan T. Barron, Sofien Bouaziz, Dan B Goldman, Ricardo Martin-Brualla, Steven M. Seitz
SIGGRAPH Asia, 2021
By applying ideas from level set methods, we can represent topologically changing scenes with NeRFs.
FiG-NeRF: Figure Ground Neural Radiance Fields for 3D Object Category Modelling
Christopher Xie, Keunhong Park, Ricardo Martin-Brualla, Matthew A. Brown
3DV, 2021
Given a lot of images of an object category, you can train a NeRF to render them from novel views and interpolate between different instances.
Nerfies: Deformable Neural Radiance Fields
Keunhong Park, Utkarsh Sinha, Jonathan T. Barron, Sofien Bouaziz, Dan B Goldman, Steven M. Seitz, Ricardo Martin-Brualla
ICCV, 2021 Oral Presentation
Learning deformation fields with a NeRF let's you reconstruct non-rigid scenes with high fidelity.
LatentFusion: End-to-End Differentiable Reconstruction and Rendering for Unseen Object Pose Estimation
Keunhong Park, Arsalan Mousavian, Yu Xiang, Dieter Fox
CVPR, 2020
By learning to predict geometry from images, you can do zero-shot pose estimation with a single network.
PhotoShape: Photorealistic Materials for Large-Scale Shape Collections
Keunhong Park, Kostas Rematas, Ali Farhadi, Steven M. Seitz
SIGGRAPH Asia, 2018 Journal Cover
By pairing large collections of images, 3D models, and materials, you can create thousands of photorealistic 3D models fully automatically.
