Yue Chen (@faneggchen) / X

Yue Chen

81 posts

Yue Chen

@faneggchen

PhD Student @Westlake_Uni. Playable world builder. Muggle dreaming of the Wizarding World

Joined December 2022

Pinned
Yue Chen
@faneggchen
Oct 8, 2025
#Human3R: Everyone Everywhere All at Once Just input a RGB video, we online reconstruct 4D humans and scene in 𝗢𝗻𝗲 model and 𝗢𝗻𝗲 stage. Training this versatile model is easier than you think – it just takes 𝗢𝗻𝗲 day using 𝗢𝗻𝗲 GPU! 🔗Page: fanegg.github.io/Human3R/
00:00
19K
Yue Chen
@faneggchen
Mar 31, 2025
How much 3D do visual foundation models (VFMs) know? Previous work requires 3D data for probing → expensive to collect! #Feat2GS @CVPR 2025 - our idea is to read out 3D Gaussains from VFMs features, thus probe 3D with novel view synthesis. 🔗Page: fanegg.github.io/Feat2GS/
00:00
24K
Yue Chen
@faneggchen
Jun 12, 2025
We're presenting Feat2GS at #CVPR2025 — probing the 3D awareness of Visual Foundation Models using only 2D images! Come chat with us to explore: 🔹 Why 2D and 3D are more aligned than you think 🔹 Which VFMs best capture geometry & appearance 🔹 How to make VFMs more 3D-aware
6.7K
Yue Chen
@faneggchen
Apr 1, 2025
Just "dissect" the cross-attention mechanism of #DUSt3R, making 4D reconstruction easier.
Xingyu Chen
@RoverXingyu
Apr 1, 2025
Replying to @RoverXingyu
💡Humans naturally separate ego-motion from object-motion without dynamic labels. We observe that #DUSt3R has implicitly learned a similar mechanism, reflected in its attention layers.
00:00
5.1K
Yue Chen
@faneggchen
Jun 15, 2025
Got you, Kosta!
Kosta Derpanis (sabbatical in Zurich)
@CSProfKGD
Jun 15, 2025
Bucket list item ✅ The Parthenon… in Nashville! 🇬🇷🎶
5.6K
Yue Chen
@faneggchen
Jun 14, 2025
Thanks for all the likes! Brilliant t-shirt idea from Yuliang
Yuliang Xiu
@yuliangxiu
Jun 14, 2025
Thanks @CVPR, for featuring our #Feat2GS T-shirts in CVPR daily magazine :)
3.8K
Yue Chen
@faneggchen
Mar 31, 2025
Replying to @faneggchen
💻Code: github.com/fanegg/Feat2GS 🎥Video: youtu.be/4fT5lzcAJqo?si… Big thanks to the amazing team! @faneggchen, @RoverXingyu, @AnpeiC, @GerardPonsMoll1, @yuliangxiu #DUSt3R #MASt3R #MiDaS #DINOv2 #DINO #SAM #CLIP #RADIO #MAE #StableDiffusion #Zero123
GitHub - fanegg/Feat2GS: [CVPR2025] Feat2GS: Probing Visual Foundation Models with Gaussian...
From github.com
4.8K
Yue Chen
@faneggchen
Mar 31, 2025
Replying to @faneggchen
With #Feat2GS we evaluated more than 10 visual foundation models (DUSt3R, DINO, MAE, SAM, CLIP, MiDas, etc) in terms of geometry and texture — see the paper for comparison. 📄Paper: arxiv.org/abs/2412.09606 🔍Try it NOW: fanegg.github.io/Feat2GS/#chart
00:00
1.4K
Yue Chen
@faneggchen
Jun 18, 2023
#CVPR2023 Hey there! Our CVPR work, UV Volumes, is now released! Any questions are welcome! fanegg.github.io/UV-Volumes UV Volumes is a deformable, real-time rendering, and editable NeRF, decomposing a dynamic human into 3D UV Volumes and a 2D appearance texture.
00:00
401
Yue Chen
@faneggchen
Apr 1, 2025
#Easi3R is a simple training-free approach adapting DUSt3R for dynamic scenes.
Xingyu Chen
@RoverXingyu
Apr 1, 2025
🦣Easi3R: 4D Reconstruction Without Training! Limited 4D datasets? Take it easy. #Easi3R adapts #DUSt3R for 4D reconstruction by disentangling and repurposing its attention maps → make 4D reconstruction easier than ever! 🔗Page: easi3r.github.io
00:00
495
Yue Chen
@faneggchen
Jun 18, 2023
#CVPR2023 Super excited to share - L2G-NeRF, our newest CVPR work, is now available!! Any questions are welcome! rover-xingyu.github.io/L2G-NeRF We combine local and global alignment via differentiable parameter estimation solvers to achieve robust bundle-adjusting Neural Fields.
00:00
405
Yue Chen
@faneggchen
Mar 31, 2025
Replying to @faneggchen
Our findings in 3D probe lead to a simple-yet-effective solution, by just combining features from different visual foundation models and outperform prior works. Apply #Feat2GS in sparse & causal captures: 🤗Online Demo: huggingface.co/spaces/endless…
00:00
854
Yue Chen
@faneggchen
Feb 27, 2025
Replying to @zhiwen_fan_ @yuliangxiu and @RoverXingyu
Thanks to InstantSplat! This work has helped us a lot!
62
Yue Chen
@faneggchen
Apr 1, 2025
Replying to @faneggchen @JulienBlanchon and @CVPR
even though DUSt3R and MASt3R are not no.1, they still made it to the top 3 – so they’re pretty good at geometry too
81