Xiuming Zhang @xiuming_zhang Twitter profile

Pinned Tweet

Xiuming Zhang

@xiuming_zhang

8 months

While you park, we 3D reconstruct. Happy holidays 🎁

Phil Duan

@philduan

8 months

happy holidays 🎁

36

34

596

25

17

380

Last Seen Profiles

@AyeDereKells

@DirtRdDemocrats

@RealCShep

@sensecapmx

@dekked_

@kkpapua1

@CcabbarovV

@lostinalone

@Sad0Bone

@KasihLudah

@DyCMGoUP

@King_feast1

@DrUnfuckable

@MineoNakab28790

@LauramcThompson

@GoddessMalika_

@PBR_MA

@marshallamps

@Nftking757064

@Riyad21350

@ezishy

@TinyFootPervert

@CollablyNetwork

@9ITCmD05e19Nr6e

@synthesischool

@samcokeradio

@ayalaa_gilda

@SeferiKeci

@Leroy_EST

@isaiah_galason

@arch_emz

@artemisvtuber

@MARCOSBARC96825

@DomusAI_com

@uscanga_pe16322

@PasutriJoss1

Xiuming Zhang

@xiuming_zhang

7 months

Real-time 3D reconstruction of "arbitrary shapes" in the wild!

Tesla

@Tesla

7 months

When parking, you can now see a high fidelity 3D representation of the world around your vehicle, including proximity & shape of nearby objects, barriers, vehicles & painted road markings By using a dedicated neural network to model obstacles & paint lines, we can accurately

934

2K

12K

14

18

326

Xiuming Zhang

@xiuming_zhang

3 years

NeRFactor is out! It's a physically-based model that factorizes appearance into shape and reflectance given just multi-view images under *one unknown* illumination (i.e., NeRF data). It supports free-viewpoint relighting (w/ shadows!) and material editing.

5

65

322

Xiuming Zhang

@xiuming_zhang

7 months

No CGI and 1x speed 🦾🤖

Tesla Optimus

@Tesla_Optimus

7 months

There’s a new bot in town 🤖 Check this out (until the very end)!

3K

7K

31K

27

18

213

Xiuming Zhang

@xiuming_zhang

3 years

🎉NeRFactor has been conditionally accepted to SIGGRAPH Asia 2021! Looking forward to presenting and chatting about it this December in Tokyo or online.

Xiuming Zhang

@xiuming_zhang

3 years

NeRFactor is out! It's a physically-based model that factorizes appearance into shape and reflectance given just multi-view images under *one unknown* illumination (i.e., NeRF data). It supports free-viewpoint relighting (w/ shadows!) and material editing.

5

65

322

0

19

172

Xiuming Zhang

@xiuming_zhang

7 months

Real-time multi-view 3D scene reconstruction in the wild 😎

Ashok Elluswamy

@aelluswamy

7 months

High-fidelity park assist is shipping this weekend to Tesla customers without ultrasonic sensors as part of the holiday release!

302

413

5K

3

9

138

Xiuming Zhang

@xiuming_zhang

3 years

📢We (Marc Levoy's computational photography team at Adobe) are looking to hire research interns for Summer 2022. Ping me if interested in working with us on relighting, neural rendering, or any computational photography problem. (Please help RT!) More info, in Marc's own words:

2

17

76

Xiuming Zhang

@xiuming_zhang

4 years

Check out our latest work, Neural Light Transport, on *simultaneous* relighting💡 and view synthesis 📷 by learning to interpolate a 6D light transport function in the texture space! Paper: Video: Project:

(OUTDATED) Neural Light Transport for Relighting and View Synthesis

THIS VERSION IS OUTDATED. PLEASE SEE THE NEWEST VERSION (https://youtu.be/gWdKjxCmYMI) INSTEAD.Project Page:http://nlt.csail.mit.edu/Authors:Xiuming Zhang, S...

www.youtube.com

Jon Barron

@jon_barron

4 years

Introducing "Neural Light Transport": Embedding a convnet within a predefined texture atlas enables *simultaneous* view synthesis and relighting, while maintaining backwards compatibility with oldschool graphics engines. Great work @xiuming_zhang ! More at

0

34

152

1

20

56

Xiuming Zhang

@xiuming_zhang

2 years

Hot take after reviewing for #CVPR2023 : IMO, we should stop calling our own methods "novel." It is subjective and carries no meaning in a scientific publication. When I was submitting to science journals, the editorial office checked for such words and asked me to remove them👍.

3

4

43

Xiuming Zhang

@xiuming_zhang

2 years

Speaking of "novel," I stopped, many years ago, using novelty as a criterion to judge papers. What's novel, and what's not? Is it a subjective or objective call? NeRF is amazing, but volume rendering, MLPs, and PE have been around. More important is whether the method works, IMO.

3

0

27

Xiuming Zhang

@xiuming_zhang

6 years

@MIT_CSAIL The recent commit should've fixed this.

0

23

Xiuming Zhang

@xiuming_zhang

1 year

Wanna build AI autonomy that runs in the real world? Come talk to us at @Tesla AI! You’ll find exciting problems to work on whether it be cars #autopilot or robots #optimus , and whether you be a high-level or mid-level vision person! #CVPR2023

Phil Duan

@philduan

1 year

@Tesla AI team is at @CVPR in Vancouver this week! If you are also here, stop by and check out what we have been working on for Autopilot, Optimus, and dojo! #CVPR2023

141

640

2K

0

1

20

Xiuming Zhang

@xiuming_zhang

4 years

Ever feeling NeRVous about tracing to every single light plus indirect illumination? NeRV can come to your rescue! It jointly optimizes for shape, light visibility, reflectance, & indirect illumination (!), while staying still tractable.

Pratul Srinivasan

@_pratul_

4 years

Check out NeRV, our latest work on recovering relightable NeRFs! @a_k_a_Billy @xiuming_zhang @BenMildenhall @jon_barron (1/3)

3

19

86

0

16

Xiuming Zhang

@xiuming_zhang

3 years

We show how to train a NeRF over a class of objects🪑💺🪑and propagate simple scribbles to 3D for color/shape editing!

Steven Liu

@stevenxliu

3 years

Our code on editing conditional radiance fields is out! We can edit the shape and color of 3D regions with simple user scribbles. With @xiuming_zhang , Zhoutong Zhang, @rzhang88 , @junyanz89 , Bryan Russell. Paper + Code + Video + Demo:

1

12

49

0

1

14

Xiuming Zhang

@xiuming_zhang

6 years

MIT news about our MoSculp work!

Creating 3-D-printed “motion sculptures” from 2-D videos

The "MoSculp" system developed at MIT CSAIL creates 3-D-printed “motion sculptures” from 2-D videos. It could help athletes, dancers, and others better analyze how they move.

news.mit.edu

1

12

Xiuming Zhang

@xiuming_zhang

2 years

Wanna stay updated on what your favorite researchers are up to?🔔Check out , which compares the webpages' current HTMLs against their previous snapshots and sends you an email of the deltas. Great for tracking folks' new papers, positions, etc. PRs welcome!

GitHub - xiumingzhang/webpage-monitor: Minimal Webpage Monitor for Academics

Minimal Webpage Monitor for Academics. Contribute to xiumingzhang/webpage-monitor development by creating an account on GitHub.

github.com

1

11

Xiuming Zhang

@xiuming_zhang

3 years

Check out the project page👆for the paper, overview video, code, and data. Joint work w/ @_pratul_ , @boyang_deng , @debfx , Bill Freeman, and @jon_barron .

1

9

Xiuming Zhang

@xiuming_zhang

8 months

Cool paper that’s so @andrewhowens 😁

Daniel Geng

@dangengdg

8 months

Can you make a jigsaw puzzle with two different solutions? Or an image that changes appearance when flipped? We can do that, and a lot more, by using diffusion models to generate optical illusions! Continue reading for more illusions and method details 🧵

16

119

609

1

0

9

Xiuming Zhang

@xiuming_zhang

8 months

I know this is supposed to be a meme now, but projective geometry 😹

International Conference on 3D Vision

@3DVconf

8 months

3D vision is nothing without _____.

50

3

42

0

2

8

Xiuming Zhang

@xiuming_zhang

11 months

Cool! "Preconditioner" sounds like camera parameter-dependent scales that "normalize" magnitudes of different camera parameters' effects so that all types of parameters receive more equal gradients during optimization. Keunhong, I would've guessed you'd name this "Equalizer." 😉

Keunhong Park

@KeunhongP

11 months

Introducing CamP🏕️ — a method to precondition camera optimization for NeRFs to significantly improve quality. With CamP we’re able to create high quality reconstructions even when input poses are bad. Project page: ArXiv: (1/n)

4

67

365

0

1

7

Xiuming Zhang

@xiuming_zhang

11 months

Cool work. The visualization reminds me of MoSculp: 🕺

Jonathon Luiten

@JonathonLuiten

11 months

Dynamic 3D Gaussians: Tracking by Persistent Dynamic View Synthesis We model the world as a set of 3D Gaussians that move & rotate over time. This extends Gaussian Splatting to dynamic scenes, with accurate novel-view synthesis and dense 3D trajectories.

25

375

2K

1

0

6

Xiuming Zhang

@xiuming_zhang

3 years

📢Consider submitting your work to the Technical Communications & Posters programs of SIGGRAPH Asia 2021, which is still scheduled to be a physical event in Tokyo🗼, with the option of online presentation if preferred -- great regardless of your travel preference!

1

4

Xiuming Zhang

@xiuming_zhang

10 months

From Bill’s “how to write a good CVPR submission” 😂

0

6

Xiuming Zhang

@xiuming_zhang

5 months

Wow! Where does the model learn the morphing effects from though? Sounds like such a niche effect buried in millions of videos.

Tim Brooks

@_tim_brooks

5 months

in addition to generating videos from text, Sora can morph between two videos. here's one example I love where it starts as a drone and turns into a butterfly

18

52

355

0

4

Xiuming Zhang

@xiuming_zhang

10 months

The demo is incredible!

Aleksander Holynski

@holynski_

10 months

Check out our new paper that turns a (single image) => (interactive dynamic scene)! I’ve had so much fun playing around with this demo. Try it out yourself on the website:

26

317

2K

0

5

Xiuming Zhang

@xiuming_zhang

7 months

@jon_barron Just like the Moon 🌒

1

0

4

Xiuming Zhang

@xiuming_zhang

4 years

Amazing speedup, and very high-quality code! Great work, @a_k_a_Billy et al.!

Jon Barron

@jon_barron

4 years

JaxNeRF! Today we're releasing Google's internal JAX implementation of NeRF. Training goes from 3 days to 2.5 hours (on a TPU pod), PSNR is slightly higher(?!), and it has all the functional/gradient goodness we love about JAX. Great work @a_k_a_Billy !

8

52

279

0

5

Xiuming Zhang

@xiuming_zhang

3 years

@docmilanfar “‘You can estimate shape, reflectance, and illumination from a single image,’ they said.“ 🤣

0

4

Xiuming Zhang

@xiuming_zhang

9 months

“Cute blue yeti driving a Model X Plaid, going fast” by DALL•E 3, which took “plaid” too literally but I like it!

0

1

4

Xiuming Zhang

@xiuming_zhang

1 year

@3DVconf Kudos to the social media chair(s) for the high-quality memes! 🥇

0

4

Xiuming Zhang

@xiuming_zhang

1 year

@jon_barron Congrats, @zhengqi_li , @QianqianWang5 , @forrestercole , Richard Tucker, and @Jimantha , for the well deserved award! Coming in with a banner like that—must’ve been confident about this happening! 🥳👏

1

0

3

Xiuming Zhang

@xiuming_zhang

4 years

@bttyeo @knutson_brain I recently adopted a pair of brothers: Scallop and Squid! :-D

1

0

3

Xiuming Zhang

@xiuming_zhang

3 years

I voted nay to this motion too because I think it's awkward to post your new paper to arXiv and meanwhile have to stay quiet about it. The ban also sounds hard to implement in practice (e.g., bot posting/tweeting is ok), although I guess most researchers will just abide by it.

Matthias Niessner

@MattNiessner

3 years

Social media is important to share research publicly and for junior researchers it's one of few ways to get yourself known. So it's not a great move to take that away when conferences are mostly virtual - I believe the CVPR motion was a big mistake and should be re-considered.

10

5

102

0

3

Xiuming Zhang

@xiuming_zhang

7 months

@jon_barron Haven’t seen any recent developments published. The latest I know of is my dissertation :)

1

0

2

Xiuming Zhang

@xiuming_zhang

3 years

@akanazawa Emacs is an OS that happens to have an editor LOL. To me, Neovim really revived Vim.

1

0

2

Xiuming Zhang

@xiuming_zhang

9 months

"Secure messages" with banks, health care providers, etc., are total scams. So annoying and not really that secure. Might as well just use email.

0

2

Xiuming Zhang

@xiuming_zhang

1 year

@YunTaTsai1 A novel physically-based face relighting method

1

0

2

Xiuming Zhang

@xiuming_zhang

11 months

@surmenok model zoo -> model arena 😉

0

2

Xiuming Zhang

@xiuming_zhang

5 months

@rzhang88 @openreviewnet Less fun to make a joke of xYuz than of R2 🤡

0

2

Xiuming Zhang

@xiuming_zhang

11 months

@jon_barron @dimadamen Spot on! People tend to call any model comprising components originally devised for real generative tasks (like diffusion models) generative. Diffusion model-based monocular depth estimation: generating depth using diffusion models (conditioning on RGB) -> generative. 😂

0

2

Xiuming Zhang

@xiuming_zhang

11 months

@jon_barron "Looping over pixels vs. primitives" -- insightful! +1 to the "neural" comment, too. Been viewing NeRF as optimization instead of learning. Then, is it, ahem, "AI"? General public probably thinks it is. Meta-NeRF sounds more like AI than vanilla NeRF. Regardless, NeRF is cool!

1

0

2

Xiuming Zhang

@xiuming_zhang

3 years

@jd_denn Yes, that's what we are trying to do, although at the current stage, "a few" may mean >50 images in most of NeRF-based approaches.

0

2

Xiuming Zhang

@xiuming_zhang

3 years

@mertrory @ngohgia Congrats, @ngohgia !

1

0

1

Xiuming Zhang

@xiuming_zhang

4 years

@simon_niklaus "show some respect"? What!!!???

1

0

1

Xiuming Zhang

@xiuming_zhang

2 years

@elliottszwu If you set the check frequency high enough (like every minute), you can indeed catch a fellow researcher fiddling with their webpage as it happens! 🕵️‍♂️👁️👀

0

1

Xiuming Zhang

@xiuming_zhang

11 months

@JonathonLuiten Very cool! Have you tried running this on another dataset?

1

0

1

Xiuming Zhang

@xiuming_zhang

3 years

@fursund @_pratul_ @boyang_deng @debfx @jon_barron That's definitely a valid extension, which we haven't had time to try out yet.

0

1

Xiuming Zhang

@xiuming_zhang

9 months

@jon_barron I told it to have opinions and not be politically correct. Then it would sometimes say “since you love opinions, blah blah” lol.

0

Xiuming Zhang

@xiuming_zhang

2 years

@jon_barron @_pratul_ @BenMildenhall Haha “exactly” 😹

0

1

Xiuming Zhang

@xiuming_zhang

7 months

@adamraudonis Root canal wtf hahaha

0

1

Xiuming Zhang

@xiuming_zhang

3 years

@taiyasaki I've been having this urge to defer it to the final section or to right before "Conclusion" for the same reasons you listed, but never ended up doing it, partly to avoid upsetting "traditional" people (who may happen to be my reviewers).

1

0

1

Xiuming Zhang

@xiuming_zhang

4 years

@jiajunwu_cs congrats!

0

1

Xiuming Zhang

@xiuming_zhang

6 years