Ben Mildenhall

I work on problems in graphics and 3D computer vision. From 2021 to 2023, I was a research scientist at Google Research. I received my PhD from UC Berkeley in 2020, where I was advised by Ren Ng and supported by a Hertz fellowship.

In the summer of 2017, I was an intern in Marc Levoy's group in Google Research. In the summer of 2018, I worked with Rodrigo Ortiz-Cayon and Abhishek Kar at Fyusion. I did my undergrad at Stanford University and worked at Pixar Research in the summer of 2014.

Email / CV / Bio / Google Scholar / Twitter / GitHub

Highlights

Neural Radiance Fields
Many images to 3D

mip-NeRF 360 / Zip-NeRF
Bigger and better NeRF

DreamFusion
Text to 3D

ReconFusion
Few images to 3D

Research

	Nuvo: Neural UV Mapping for Unruly 3D Representations Pratul Srinivasan, Stephan J. Garbin, Dor Verbin, Jonathan T. Barron, Ben Mildenhall arXiv, 2023 project page / video / arXiv Use neural fields to recover editable UV mappings for challenging geometry (e.g. NeRFs, marching cubes meshes, DreamFusion).
	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* arXiv, 2023 project page / arXiv Finetune an image diffusion model to accept multiview inputs, then use it to regularize radiance field reconstruction.
	Generative Powers of Ten Xiaojuan Wang, Janne Kontkanen, Brian Curless, Steve Seitz, Ira Kemelmacher, Ben Mildenhall, Pratul Srinivasan, Dor Verbin, Aleksander Holynski arXiv, 2023 project page / arXiv Use a text-to-image model to generate consistent content across drastically varying scales.
	CamP: Camera Preconditioning for Neural Radiance Fields Keunhong Park, Philipp Henzler, Ben Mildenhall, Jonathan T. Barron, Ricardo Martin-Brualla SIGGRAPH Asia, 2023 project page / arXiv Preconditioning based on camera parameterization helps NeRF and camera extrinsics/intrinsics optimize better together.
	Zip-NeRF: Anti-Aliased Grid-Based Neural Radiance Fields Jonathan T. Barron, Ben Mildenhall, Dor Verbin, Pratul Srinivasan, Peter Hedman ICCV, 2023 (Best Paper Finalist) project page / video / arXiv Combining mip-NeRF 360 and Instant NGP lets us reconstruct huge scenes.
	DreamBooth3D: Subject-Driven Text-to-3D Generation Amit Raj, Srinivas Kaza, Ben Poole, Michael Niemeyer, Nataniel Ruiz, Ben Mildenhall, Shiran Zada, Kfir Aberman, Michael Rubinstein, Jonathan T. Barron, Yuanzhen Li, Varun Jampani ICCV, 2023 project page / arXiv Combining DreamBooth (personalized text-to-image) and DreamFusion (text-to-3D) yields high-quality, subject-specific 3D assets with text-driven modifications
	Eclipse: Disambiguating Illumination and Materials using Unintended Shadows Dor Verbin, Ben Mildenhall, Peter Hedman, Jonathan T. Barron, Todd Zickler, Pratul Srinivasan arXiv, 2023 project page / video / arXiv Shadows cast by unobserved occluders provide a high-frequency cue for recovering illumination and materials.
	BakedSDF: Meshing Neural SDFs for Real-Time View Synthesis Lior Yariv, Peter Hedman, Christian Reiser, Dor Verbin, Pratul Srinivasan, Richard Szeliski, Jonathan T. Barron, Ben Mildenhall SIGGRAPH, 2023 project page / video / arXiv We achieve real-time view synthesis by baking a high quality mesh and fine-tuning a lightweight appearance model on top.
	MERF: Memory-Efficient Radiance Fields for Real-time View Synthesis in Unbounded Scenes Christian Reiser, Richard Szeliski, Dor Verbin, Pratul Srinivasan, Ben Mildenhall, Andreas Geiger, Jonathan T. Barron, Peter Hedman SIGGRAPH, 2023 project page / video / arXiv We achieve real-time view synthesis using a volumetric rendering model with a compact representation combining a sparse 3D feature grid and 2D feature planes.
	AligNeRF: High-Fidelity Neural Radiance Fields via Alignment-Aware Training Yifan Jiang, Peter Hedman, Ben Mildenhall, Dejia Xu, Jonathan T. Barron, Zhangyang Wang, Tianfan Xue CVPR, 2023 project page / arXiv Accounting for misalignment due to scene motion or calibration errors improves NeRF reconstruction quality.
	DreamFusion: Text-to-3D using 2D Diffusion Ben Poole, Ajay Jain, Jonathan T. Barron, Ben Mildenhall ICLR, 2023 (Outstanding Paper Award) project page / arXiv / gallery We optimize a NeRF from scratch using a pretrained text-to-image diffusion model to do text-to-3D generative modeling.
	Fast and High-quality Image Denoising via Malleable Convolutions Yifan Jiang, Bartlomiej Wronski, Ben Mildenhall, Jonathan T. Barron, Zhangyang Wang, Tianfan Xue ECCV, 2022 project page / arXiv We denoise images efficiently by predicting spatially-varying kernels at low resolution and using a fast fused op to jointly upsample and apply these kernels at full resolution.
	NeRF in the Dark: High Dynamic Range View Synthesis from Noisy Raw Images Ben Mildenhall, Peter Hedman, Ricardo Martin-Brualla, Pratul Srinivasan, Jonathan Barron CVPR, 2022 (oral) project page / arXiv / video / code We train RawNeRF directly on linear raw camera images, enabling new HDR view synthesis applications and greatly increasing robustness to camera noise.
	Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields Jonathan T. Barron, Ben Mildenhall, Dor Verbin, Pratul Srinivasan, Peter Hedman CVPR, 2022 (oral) project page / arXiv / video / code We extend mip-NeRF to produce photorealistic results on unbounded scenes.
	Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields Dor Verbin, Peter Hedman, Ben Mildenhall, Todd Zickler, Jonathan T. Barron, Pratul Srinivasan CVPR, 2022 (Best Student Paper Honorable Mention) project page / arXiv / video / code Explicitly modeling reflections in NeRF produces realistic shiny surfaces and accurate surface normals, and lets you edit materials.
	Block-NeRF: Scalable Large Scene Neural View Synthesis Matthew Tancik, Vincent Casser, Xinchen Yan, Sabeek Pradhan, Ben Mildenhall, Pratul Srinivasan, Jonathan T. Barron, Henrik Kretzschmar CVPR, 2022 (oral) project page / arXiv / video We build city-scale scenes from many NeRFs, trained using millions of images.
	RegNeRF: Regularizing Neural Radiance Fields for View Synthesis from Sparse Inputs Michael Niemeyer, Jonathan T. Barron, Ben Mildenhall, Mehdi S. M. Sajjadi, Andreas Geiger, Noha Radwan CVPR, 2022 (oral) project page / arXiv / video We regularize unseen views during optimization to enable view synthesis from as few as 3 input images.
	Zero-Shot Text-Guided Object Generation with Dream Fields Ajay Jain, Ben Mildenhall, Jonathan T. Barron, Pieter Abbeel, Ben Poole CVPR, 2022 project page / arXiv / video / code Supervising the CLIP embeddings of NeRF renderings allows us to generate 3D objects from text prompts alone.
	Dense Depth Priors for Neural Radiance Fields from Sparse Input Views Barbara Roessle, Jonathan T. Barron, Ben Mildenhall, Pratul Srinivasan, Matthias Nießner CVPR, 2022 arXiv / video We apply dense depth completion techniques to freely-available sparse stereo data to guide NeRF reconstructions from few input images.
	Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields Jonathan T. Barron, Ben Mildenhall, Matthew Tancik, Peter Hedman, Ricardo Martin-Brualla, Pratul Srinivasan ICCV, 2021 (Best Paper Honorable Mention) project page / arXiv / video / code We prefilter the positional encoding function and train NeRF to generate anti-aliased renderings.
	Baking Neural Radiance Fields for Real-Time View Synthesis Peter Hedman, Pratul Srinivasan, Ben Mildenhall, Jonathan T. Barron, Paul Debevec ICCV, 2021 (oral) project page / arXiv / video / demo We bake a trained NeRF into a sparse voxel grid of colors and features in order to render it in real-time.
	Learned Initializations for Optimizing Coordinate-Based Neural Representations Matthew Tancik, Ben Mildenhall, Terrance Wang, Divi Schmidt, Pratul Srinivasan, Jonathan T. Barron, Ren Ng CVPR, 2021 (oral) project page / arXiv / video / code We use meta-learning to find weight initializations for coordinate-based MLPs that allow them to converge faster and generalize better.
	NeRV: Neural Reflectance and Visibility Fields for Relighting and View Synthesis Pratul Srinivasan, Boyang Deng, Xiuming Zhang, Matthew Tancik, Ben Mildenhall, Jonathan T. Barron CVPR, 2021 project page / arXiv / video We recover relightable NeRF-like models using neural approximations of expensive visibility integrals, so we can simulate complex volumetric light transport during training.
	Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains Matthew Tancik, Pratul Srinivasan, Ben Mildenhall, Sara Fridovich-Keil, Nithin Raghavan, Utkarsh Singhal, Ravi Ramamoorthi, Jonathan T. Barron, Ren Ng NeurIPS, 2020 (spotlight)* project page / arXiv / code We demonstrate that composing fully-connected networks with a simple Fourier feature mapping allows them to learn much high frequency functions.
	Neural Reflectance Fields for Appearance Acquisition Sai Bi, Zexiang Xu, Pratul Srinivasan, Ben Mildenhall, Kalyan Sunkavalli, Milos Hasan, Yannick Hold-Geoffroy, David Kriegman, Ravi Ramamoorthi arXiv, 2020 arXiv We recover relightable NeRF-like models by predicting per-location BRDFs and surface normals, and marching light rays through the NeRV volume to compute visibility.
	NeRF: Representing Scenes as Neural Radiance Fields for View Synthesis Ben Mildenhall, Pratul Srinivasan, Matthew Tancik, Jonathan T. Barron, Ravi Ramamoorthi, Ren Ng ECCV, 2020 (Best Paper Honorable Mention)* project page / arXiv / video / talk / code / two minute papers / papers with code / wired We optimize a simple fully-connected network to represent a single scene as a volume, then use volume rendering to do view synthesis.
	Deep Multi Depth Panoramas for View Synthesis Kai-En Lin, Zexiang Xu, Ben Mildenhall, Pratul Srinivasan, Yannick Hold-Geoffroy, Stephen DiVerdi, Qi Sun, Kalyan Sunkavalli, Ravi Ramamoorthi ECCV, 2020 arXiv / video We represent scenes as multi-layer panoramas with depth for VR view synthesis.
	Lighthouse: Predicting Lighting Volumes for Spatially-Coherent Illumination Pratul Srinivasan, Ben Mildenhall, Matthew Tancik, Jonathan T. Barron, Richard Tucker, Noah Snavely CVPR, 2020 project page / arXiv / video / code We predict a volume from an input stereo pair that can be used to calculate incident lighting at any 3D point within a scene.
	StegaStamp: Invisible Hyperlinks in Physical Photographs Matthew Tancik, Ben Mildenhall, Ren Ng CVPR, 2020 project page / arXiv / video / code We can hide hyperlinks in natural images to create aesthetically pleasing barcodes.
	Local Light Field Fusion: Practical View Synthesis with Prescriptive Sampling Guidelines Ben Mildenhall, Pratul Srinivasan, Rodrigo Ortiz-Cayon, Nima Khademi Kalantari, Ravi Ramamoorthi, Ren Ng, Abhishek Kar SIGGRAPH, 2019 project page / arXiv / video / code We develop and analyze a deep learning method for rendering novel views of complex real world scenes.
	Unprocessing Images for Learned Raw Denoising Tim Brooks, Ben Mildenhall, Tianfan Xue, Jiawen Chen, Dillon Sharlet, Jonathan T. Barron CVPR, 2019 (oral) project page / arXiv We can learn a better denoising model by processing and unprocessing images the same way a camera does.
	Burst Denoising with Kernel Prediction Networks Ben Mildenhall, Jonathan T. Barron, Jiawen Chen, Dillon Sharlet, Ren Ng, Robert Carroll CVPR, 2018 (spotlight) project page / arXiv / code We train a network to predict linear kernels that denoise bursts of raw linear images.
	DiffuserCam: Lensless Single-exposure 3D Imaging Nick Antipa, Grace Kuo, Reinhard Heckel, Ben Mildenhall, Emrah Bostan, Ren Ng, Laura Waller Optica, 2018 project page / arXiv Using a diffuser instead of a lens lets you recover 3D in a single exposure.
	Approximations for the distribution of microflake normals Nelson Max, Tom Duff, Ben Mildenhall, Yajie Yan The Visual Computer, 2017 We precompute microflake approximations to make rendering large meshes at a distance more efficient.
	Controlling Procedural Modeling Programs with Stochastically-Ordered Sequential Monte Carlo Daniel Ritchie, Ben Mildenhall, Noah D. Goodman, Pat Hanrahan SIGGRAPH, 2015 We improve control over the output of highly-variable procedural modeling programs by using SOSMC to provide incremental feedback on partially-generated models.