Shuzhe Wang

Researcher @ Apple, PhD from Aalto University

[Shuzhe Wang @riverakid1]

Thank you so much, Jianyuan @jianyuan_wang ! Big congrats on the Best Paper - VGGT is well deserved and truly inspiring! 🏆 I really appreciate the kind words about DUSt3R! It’s been a pleasure getting to know you these past months - excited to see where your research goes next🚀

[Jianyuan @jianyuan_wang]

Just touched down from the air back to solid ground😄 We're honored to receive the Best Paper Award this year, while I want to give a special shout-out to the authors of the brilliant paper Dust3R: Shuzhe Wang (@riverakid1), Vincent Leroy (@Vinc3nt_Leroy), Yohann Cabon, Boris Chidlovskii, and Jerome Revaud (@JeromeRevaud).

It is Dust3R that first showcased the power of scaling up 3D learning (by aggregating academic datasets) and trusting data! This paved the way for our research🥰 On a personal note, I only met Shuzhe Wang (@riverakid1) two months ago, but it was immediately clear he's both insightful and genuinely awesome 😊 Any team would be incredibly lucky to hire him!

Finally, huge thanks to all—old friends and new—for an amazing CVPR experience.

[Shuzhe Wang @riverakid1]

With @i411648085, @B1ueber2y, Lulu Cai, @FanQingnan , Juho Kannala, Yanchao Yang

[Shuzhe Wang @riverakid1]

DUSt3R meets pose regression! 🌟 Introducing Reloc3r: a generalizable, fast, and accurate framework for visual localization.
Combining ViTs with motion averaging, Reloc3r sets a strong baseline for relative pose regression.

Check out the code & paper: https://t.co/NHxCN74Fbr 🎯

[Shuzhe Wang @riverakid1]

With Yuzheng Liu, @i411648085, Yanchao Yang, @FanQingnan, Baoquan Chen

[Shuzhe Wang @riverakid1]

🚀 Excited to introduce SLAM3R, a simple and effective dense scene reconstruction system for monocular RGB videos.

On top of DUSt3r, SLAM3R provides:
✅ Real-time performance
✅ High-quality reconstruction
✅ Pose-free estimation

Code available: https://t.co/CpbjH9R19W

[Vincent Leroy @Vinc3nt_Leroy]

Let's all welcome MASt3R, the new *St3R model that just came to this world! My laptops Quadro RTX 3000 (6GB) can process ~100 images in 10 min, that's pretty nice for mapping!

[Vincent Leroy @Vinc3nt_Leroy]

Awesome inspiring talk from @JeromeRevaud !!! So delighted by the positive response of everyone about our work for the past 4 years ! And what a crowd!

[Shuzhe Wang @riverakid1]

@JeromeRevaud @WayneINR @AjdDavison @LourdesAgapito The mast3r results are awesome, huge improvement over dust3r! It seems that we find the right direction for data-driven 3D reconstruction.

[Rui Li @leedaray]

🚀Excited to introduce the open-source 🌟Awesome-DUSt3R🌟, a curated list of DUSt3R-related (https://t.co/pUD8YGbIYw) 📄papers, 📚blog posts, 🎥videos, etc. Join us on this journey and contribute to the ever-growing list at https://t.co/Iwx79vjlEw!

[hu-po @hupobuboo]

better faster stronger 3DGS, thanks @janusch_patas for the recommendation https://t.co/IM5CGX6E9R

[Shuzhe Wang @riverakid1]

The bar is too high to reach…😕

[Jerome Revaud @JeromeRevaud]

our #DUSt3R is not even a spotlight, it's a poster. Even though we had excellent reviews... I don't get it

[Arno Solin @arnosolin]

📣We are hiring in 🇫🇮!
We are looking for motivated applicants to do a PhD in AI/ML-related topics. Please, spread the word.

[Jerome Revaud @JeromeRevaud]

Some more impressive 3D reconstruction examples from #DUSt3R
kindly prepared by our boss ;-) @martinhu @naverlabseurope

[Jerome Revaud @JeromeRevaud]

Sth I find fascinating with DUSt3R is its uncanny ability to perform auto-calibration... even with single images. Is it precise? Quite so, as you can see below.

This challenges the prevailing notion that single-image calibration is ill-posed (ambiguous & under-constrained)!

[Jerome Revaud @JeromeRevaud]

done, DUSt3R now first of the leaderboard!
https://t.co/I0p6HHykVL

[cocktail peanut @cocktailpeanut]

Instant Video-to-3D with DUST3R

Dust3r generates a whole 3D scene from just a couple of images. What if it could:

1. Accept a VIDEO
2. Extract the video frames
3. Turn them into 3D?

So added a Gradio Video Component.

Here's me generating 3D from a video cc: @naverlabseurope

[Jerome Revaud @JeromeRevaud]

Another convenient usage of DUSt3R (kind of anecdotal):
While looking for a rental apartment on airbnb for my vacations, i noticed I had difficulty grasping the layout and space of the apartment based on photos alone.
Solution: put all photos in DUSt3R and voila :)

[Chris Offner @chrisoffner3d]

@riverakid1 @JeromeRevaud @bchidlovskii @Vinc3nt_Leroy Witchcraft. 😄 I just ran it on my own images (after changing all mentions of CUDA to [device = "mps" if https://t.co/8EfxodAoLJ_available() else "cpu"]) and wow, this is super cool! 🤩

[Chris Offner @chrisoffner3d]

@JeromeRevaud 😳

[Vincent Leroy @Vinc3nt_Leroy]

Been waiting for the code release of DUSt3R? Wait no more, there it is! https://t.co/tN7R1Xa8qQ
Reminiscent of RayDiffusion (https://t.co/UaQrTSVAGO), we present a well behaved variant to Bundle Adjustment, this time with pointmaps.