Ang Cao

I am a final-year Ph.D. student (2020-) in Computer Science and Engineering (CSE) at the University of Michigan, Ann Arbor, supervised by Prof. Justin Johnson and Prof. JJ (Jeong Joon) Park. I was very fortunate to be advised by Prof. Andrew Owens. I work on 3D vision: primarily how to represent, understand and generate the 3D/4D world.

Before that, I was a M.S. Student at UMich ECE (2018-2020) and I did my Bachelor's degree at Wuhan University in China (2014-2018).

I am actively looking for full-time/postdoc positions in 2025!

Email / CV / Google Scholar / Twitter / Github

Work Experience

Research Scientist Intern, Meta FAIR, MPK, USA

Work with Sasha Sax , Franziska Meier

May 2024 - December 2024

Research Scientist Intern, Meta GenAI, London, UK

Work with David Novotny, Andrea Vedaldi , Natalia Neverova

May 2023 - November 2023

Publications

Equal contribution is indicated by * and listed in alphabetical order.

	LIFT-GS: Cross-Scene Render-Supervised Distillation for 3D Language Grounding Ang Cao, Sergio Arnaud, ... Justin Johnson, JJ Park, Alexander (Sasha) Sax ArXiv, 2025 ArXiv/ website We (pre-)train a 3D Visual Language Grounding (3D VLG) model with only 2D supervision, by distilling languages from 2D foundation models with render-supervision.
	Probing Visual Language Priors in VLMs Ang Cao, Tiange Luo, Gunhee Lee Justin Johnson, Honglak Lee ArXiv, 2025 ArXiv/ How could a generator help the agent? We explore the visual language priors in VLMs by constructing novel question-image-answer triplets from image diffusion models. Proposed Image-DPO to encourage the model to use more visual inputs.
	Fast3R: Towards 3D Reconstruction of 1000+ Images in One Forward Pass Jianing "Jed" Yang, Alexander Sax, Kevin J. Liang, Mikael Henaff, Hao Tang, Ang Cao, Joyce Chai, Franziska Meier Matt Feiszli CVPR, 2025 ArXiv/ website/ code/ Demo/ We present a neat transformer-based 3D recon and camera pose est. pipeline, which can reconstruct 3D scenes from 1000+ images in one forward pass with ultra-high speed.
	Meta 3D Gen Meta GenAI 3DGen Team Tech report, 2024 Paper/ A new state-of-the-art, fast pipeline for text-to-3D asset generation
	Lightplane: Highly-Scalable Components for Neural 3D Fields Ang Cao, Justin Johnson, Andrea Vedaldi, David Novotny 3DV, 2025 project page/ ArXiv/ Docs/ code We investigate "flashattention" for NeRF: we present Lightplane Splatter and Lightplane Renderer, a pair of extremely memory efficient modules which can lift 2D images to 3D and render from theoretically any 3D hash representation with 4-5 orders of magnitude memory savings. We show its usage in a set of 3D recon and generation tasks.
	DreamGaussian4D: Generative 4D Gaussian Splatting Jiawei Ren, Liang Pan, Jiaxiang Tang , Chi Zhang, Ang Cao, Gang Zeng, Ziwei Liu, ArXiv, 2023 ArXiv/ Website/ Code/ We do 4D generation with Gaussian Splatting by distilling motions from video diffusion models.
	Text2Room: Extracting Textured 3D Meshes from 2D Text-to-Image Models Ang Cao , Lukas Höllein , Andrew Owens, Justin Johnson, Matthias Nießner ICCV, Oral , 2023 project page/ ArXiv/ video/ code We generate meshes of full 3D rooms using text-to-image models.
	HexPlane: A Fast Representation for Dynamic Scenes Ang Cao, Justin Johnson CVPR, 2023 project page/ ArXiv/ video/ code An elegant representation for dynamic 3D scenes using six feature planes.
	FWD: Real-time Novel View Synthesis with Forward Warping and Depth Ang Cao, Chris Rockwell, Justin Johnson CVPR, 2022 project page / ArXiv / video / code / We show point rasterization can be really fast for sparse view novel view synthesis.
	Inverting and Understanding Object Detectors Ang Cao, Justin Johnson Tech report, 2021 ArXiv / code / Revealing intriguing properties of detectors by applying our layout inversion technique.

Teaching Experience

EECS 498/598: Computer Graphics and Generative Models (Fall 2024)

Teaching assistant (GSI/Head TA), working with JJ Park

Huge thanks to Jon Barron for the awesome template.