Roy Yang
@YungHsuYang
Followers
90
Following
427
Media
3
Statuses
106
Ph.D. student @ ETH Zürich
Zürich, Switzerland
Joined June 2017
Very excited to open the afternoon Oral session with SuperDec 🏝️🚀 ✨ Oral: 1PM @ Exhibit Hall III (Ground Floor) 🖼️ Poster: 2:30PM, #143 🌍: https://t.co/H1CmUxp8Wc
@ICCVConference @mapo1 @FrancisEngelman @BoyangBoysun
Are photorealistic representations all we need? In SuperDec, we turn millions of points into compact and modular abstractions made of just a few superquadrics!🧩 Try our code and get a compact representation of your favorite scene!🚀 👾: https://t.co/4MSfLuUlTl
1
4
25
Excited to present FlowR @ICCVConference ! Come by our highlight poster this afternoon at 2:30PM, booth 288, Exhibit Hall II to talk generative models + 3D reconstruction 🫶
1/3 Introducing FlowR 🌸: Flowing from Sparse to Dense 3D Reconstructions We learn a direct mapping between incorrect renderings and their corresponding ground-truth images, augmenting scene captures with consistent novel, generated views to improve reconstruction quality.
0
7
64
See you today at #ICCV2025 this afternoon 14:30-16:30 at poster 301!
We will present QuickSplat at #ICCV2025! 🎉 Data-driven 2DGS initialization and densification makes 3D surface reconstruction fast & accurate! 🚀 Projcet: https://t.co/SKJ7XRWTVt Arxiv:
0
3
9
We are ready for the second day of #ICCV2025 🌺 you can find us at: 🐊 CroCoDL Workshop: 301B, 1:00PM ➡️ https://t.co/c1reN3cgJb ✨RANSAC Tutorial: 328, 1:00PM ➡️ https://t.co/ltPMZE9rqn
@zuriich92 @hermannsblum @majti89 @mapo1 @ICCVConference
0
8
15
#ICCV2025 is starting today 🌺 you can find us at: 🥽 LaMAria Tutorial: 325B, 1:00PM ➡️ https://t.co/X9UBfzNGm1 ☀️ OpenSUN3D Workshop: 306 B, 1:30PM ➡️ https://t.co/wmy9HtP7dw
@pesarlin @efedele16 @FrancisEngelman @AlexDelitzas @mapo1 @ICCVConference
0
8
14
Let your robot peek around corners, size up the gap between chairs, and know exactly where everything sits. 🤖 Our new work SR-3D masters this kind of spatial reasoning. It learns from multiple views to understand distances, layouts, and how objects relate in 3D space.
6
39
163
Our new benchmark iPhone NVS. Super interesting when the data captures are not “perfect”.
📢📢📢We've released the ScanNet++ Novel View Synthesis Benchmark for iPhone data! 🥳 Test your models on RGBD video featuring real-world challenges like exposure changes & motion blur! Download the newest iPhone NVS test split and submit your results! ⬇️ https://t.co/hLnFwifvTL
0
7
67
Are photorealistic representations all we need? In SuperDec, we turn millions of points into compact and modular abstractions made of just a few superquadrics!🧩 Try our code and get a compact representation of your favorite scene!🚀 👾: https://t.co/4MSfLuUlTl
8
30
204
🚀Excited to share our recent work on test-time scaling for feed-forward Gaussian splatting: we learn a recurrent model ReSplat that is able to iteratively improve the reconstruction quality in a feed-forward manner! https://t.co/OB38xC7PjC
5
49
306
1/4 🚀 We’re excited to release MVTracker (ICCV 2025 Oral), the first data-driven multi-view 3D point tracker. MVTracker tracks arbitrary 3D points across multiple cameras, handling occlusions and varied camera setups without per-sequence optimization.
12
49
333
Check our website for more details: https://t.co/QdPXeD0C4b And try our demo with your own images!
0
2
3
Excited to share our new work for monocular open-set 3D object detection! We propose the first end-to-end 3D Monocular Open-set Object Detector (3D-MOOD), via 2D to 3D lifting, geometry-aware 3D query generation, and canonical image space. #ICCV2025 Code: https://t.co/tLB2XFOUnr
2
1
4
We will present QuickSplat at #ICCV2025! 🎉 Data-driven 2DGS initialization and densification makes 3D surface reconstruction fast & accurate! 🚀 Projcet: https://t.co/SKJ7XRWTVt Arxiv:
arxiv.org
Surface reconstruction is fundamental to computer vision and graphics, enabling applications in 3D modeling, mixed reality, robotics, and more. Existing approaches based on volumetric rendering...
📢 QuickSplat: Fast 3D Surface Reconstruction via Learned Gaussian Initialization @liuyuehcheng learns 2DGS initialization, densification, and optimization priors from ScanNet++ => fast & accurate reconstruction! Project: https://t.co/mDgQxmhqkF
0
20
110
All the talks from our @CVPR workshop on Open-World 3D Scene Understanding with Foundation Models are now available on Youtube -- enjoy! https://t.co/5id1OYw68w
youtube.com
The recordings from our workshop on Open-World 3D Scene Understanding @CVPR are now available! See you @ICCVConference in Honolulu🏄♂️ for the next edition! ➡️ https://t.co/bMc47CMdVR 🌍 https://t.co/4nSXaJHlfp
1
13
97
📢 Introducing GenSfM – a generic Structure-from-Motion pipeline built on COLMAP, with smooth, invertible, and universal calibration for diverse cameras — from pinhole to catadioptric! Thanks @don2qian @visionviktor @mapo1 for the great collaboration. 🔗 https://t.co/Wa30YDf4vb
2
36
143
📢 The first X-Sense Workshop: Ego-Exo Sensing for Smart Mobility at #ICCV2025! 🎤 We’re honored to host an outstanding speaker lineup, featuring Manmohan Chandraker, @BharathHarihar3, @wucathy, Holger Caesar, @zhoubolei, @Boyiliee, Katie Luo https://t.co/FmVGnwv906
1
16
24
🌺 Excited to announce the 1st CroCoDL Workshop on Large-Scale Cross-Device Localization at #ICCV2025! Tackling real-world localization across smartphones, AR/VR, and robots with invited talks, paper track & competition on a novel dataset. 🔗 https://t.co/E41dLsdFvl 🧵 1/2
2
5
16