峥 曾<br>Zheng Zeng

｢世界中の名もなき光｣
｢繋がっていく 星座のように｣

About

Hi! I'm Zheng, a fourth-year Ph.D. candidate at UC Santa Barbara, working with Prof. Lingqi Yan. I earned my Master's and Bachelor's degrees from Shandong University, working with Professors Lu Wang, Yanning Xu, and remotely with Prof. Beibei Wang.

I've interned at NVIDIA with Edward Liu and at Adobe with Miloš Hašan, Valentin Deschaintre, Iliyan Georgiev, Yannick Hold-Geoffroy, Yiwei Hu, and Fujun Luan.

My research focuses on achieving real-time path tracing with ultimate photorealism, exploring geometric representations, sampling methods, complex appearance, dynamic global illumination solutions, software ray tracing solutions, and post-processing techniques.

I also study emerging technologies to enhance real-time rendering, such as compact neural networks to compress/bake code and data for efficiency and generative models for photorealism.

Publications

rgbx24 publication teaser
RGB↔X: Image Decomposition and Synthesis Using Material- and Lighting-aware Diffusion Models
"You Can Let GenAI Models Eliminate/Simulate Global Illumination"
Zheng Zeng, Valentin Deschaintre, Iliyan Georgiev, Yannick Hold-Geoffroy, Yiwei Hu, Fujun Luan, Ling-Qi Yan, Miloš Hašan
ACM SIGGRAPH 2024 (Conference Track)

Project Page Paper Code Slides Supplementary DOI arXiv

extrass23 publication teaser
ExtraSS: A Framework for Joint Spatial Super Sampling and Frame Extrapolation
"You Can Jointly Upsample and Extrapolate Frames Within One Framework"
Songyin Wu, Sungye Kim, Zheng Zeng, Deepak Vembar, Sangeeta Jha, Anton Kaplanyan, Ling-Qi Yan
ACM SIGGRAPH Asia 2023 (Conference Track)

Project Page Paper Supplementary Supplemental video DOI

roma23 publication teaser
Ray-aligned Occupancy Map Array for Fast Approximate Ray Tracing
"You Can Voxelize Objects From Different Directions to Minimize Execution Divergence"
Zheng Zeng, Zilin Xu, Lu Wang, Lifan Wu, Ling-Qi Yan
Eurographics Symposium on Rendering 2023 (CGF track)

Project Page Paper Slides Supplemental video DOI Code (coming soon)

neuralbasis22 publication teaser
Lightweight Neural Basis Functions for All-Frequency Shading
"You Can Have Neural Basis Functions Act as Real Ones"
Zilin Xu, Zheng Zeng, Lifan Wu, Lu Wang, Ling-Qi Yan
SIGGRAPH ASIA 2022

Project Page Paper Supplementary DOI

rtg221 publication teaser
Temporally Reliable Motion Vectors for Better Use of Temporal Information
Zheng Zeng, Shiqiu (Edward) Liu, Jinglei Yang, Lu Wang, Ling-Qi Yan
Ray Tracing Gems II

Project Page

trmv21 publication teaser
Temporally Reliable Motion Vectors for Real-time Ray Tracing
"You Can Design Different Motion Vectors for Different Effects to Better Utilize Temporal Information"
Zheng Zeng, Shiqiu (Edward) Liu, Jinglei Yang, Lu Wang, Ling-Qi Yan
Computer Graphics Forum (Proceedings of Eurographics 2021)

Project Page Paper Supplemental video DOI

svbrdfgan20 publication teaser
Joint SVBRDF Recovery and Synthesis From a Single Image using an Unsupervised Generative Adversarial Network
"You Don't Need Corresponding SVBRDFs to Train a GAN to Generate SVBRDFs"
Yezi Zhao, Beibei Wang, Yanning Xu, Zheng Zeng, Lu Wang, Nicolas Holzschuch
Eurographics Symposium on Rendering 2020

Project Page Paper Code DOI

sppmdenoiser20 publication teaser
Denoising Stochastic Progressive Photon Mapping Renderings Using a Multi-Residual Network
"You Need Different Receptive Fields for Different Scales of Noise"
Zheng Zeng, Lu Wang, Beibei Wang, Chun-Meng Kang, Yanning Xu
Journal of Computer Science and Technology (CVM2020)

Project Page Paper Slides DOI

Misc

I enjoy optimizing my research workflow. Tools I find useful:

  • Read papers, make sense of research, and develop intuition: Heptabase.
    • My path on note-taking system: markdown + git → Notion → Obsidian → Logseq → Roam Research → Heptabase.
    • My personal opinion: note-taking is all about spatial (across different projects) and temporal (from past and for future) knowledge reuse.
  • Find related papers: Connected Papers.

I collect research self-development resources in teach-yourself-research.

Here are two renderings I made while I was taking Wei Dai’s online course; I started my rendering journey since then on.

My interests and favorites:

  • J-pop: songs from 家入レオ (推し), Official Hige Dandism, Back Number, Milet, Creepy Nuts, Tokyo Jihen, Laika, m-flo
  • Manzai (漫才): from Non style, 令和ロマン, Sandwichman, Miki, Knights, Tom Brown
  • Films and TV series: 猫の恩返し, 千と千尋の神隠し, Interstellar, The Lord of the Rings series, Ready Player One, リトル・フォレスト, ゴールデンスランバー, Look Back, 三国演义 (1994), 古畑任三郎, ブラッシュアップライフ, 重版出来, 白い巨塔, La Grande Maison Tokyo
  • Video games: Diablo II, Starcraft II, Warcraft III, Dota 2, Command & Conquer: Red Alert 3, Grim Dawn, Age of Empires IV
  • Anime: JoJo’s Bizarre Adventure, シロバコ
  • Tokusatsu (特撮), Manga, Photography

Quotes:

  • “目指して欲しいのは、人が見ていたら努力だと思うけど、本人はそう思っていない状態。私にとっては普通の事なのに、なんか周りは評価してくれる、だったらすごくいい状態。” - 鈴木 一朗
  • “やってみて「ダメだ」と、わかった事とはじめから「ダメだ」と、言われたことは、違います。” - 鈴木 一朗
  • “世の中の大事なことって、たいてい面倒くさいんだよ。” - 宮崎駿
  • “Research is to see what everybody has seen and think what nobody has thought.” – Albert Szent-Györgyi paraphrasing Schopenhauer
  • “It’s hard to do a really good job on anything you don’t think about in the shower.” - Paul Graham
  • “Try to learn something about everything and everything about something.” ― Thomas Henry Huxley