Jiahui
@JiahuiLei1998
Followers
702
Following
310
Media
9
Statuses
60
CS PhD student @Penn
Philadelphia
Joined February 2019
Don't wanna be here?
Send us removal request.
Explore trending content on Musk Viewer
Brazil
• 1338090 Tweets
#BUS1stFANCON_KnockKnockKnock
• 364381 Tweets
アニサマ
• 192122 Tweets
GULF NEW SONG
• 67175 Tweets
TEDDYTINT x POND
• 49575 Tweets
Djokovic
• 29719 Tweets
음악중심
• 28521 Tweets
札幌2歳
• 25415 Tweets
Briggitte
• 24639 Tweets
Ticketmaster
• 22238 Tweets
JAMESSU FM IN QINGDAO
• 21275 Tweets
Stanford
• 17564 Tweets
Popyrin
• 17358 Tweets
SUNSET X VIBES WATCH PARTY
• 15011 Tweets
레이브 1위
• 14330 Tweets
台風一過
• 13956 Tweets
メドゥーサ
• 10137 Tweets
A tiny step towards modeling the dynamic real world.
Given a casual monocular video, MoSca reconstructs a renderable 4D scene. It provides a global reconstruction and is COLMAP-Free. Two videos below are from SORA.
Project:
Video:
3
23
156
Our wonderful master students
@TimSong52005757
and his classmates re-implemented the 2DGS! a bold step towards surfels and geometry reconstruction from Gaussian Splatting!
2
17
133
Happy to introduce our new
#CVPR2023
paper EFEM with
@CongyueD
, Karl Schmeckpeper ,
@GuibasLeonidas
,
@KostasPenn
:
- Learn Equivariant object shape prior on ShapeNet;
- Directly Inference for scene object segmentation!
- A new dataset "Chairs and Mugs"!
2
12
65
Our Equivariance workshop [
@KostasPenn
@CongyueD
@xu_yinshua86846
@kiamada
@ericyi0124
and Christine Allen-Blanchette] has been accepted to CVPR 2024! See you in Seattle!
0
4
49
A nice following-up combing our 2022 CVPR … CaDeX with Neural Rendering for consistent tracking
Tracking Everything Everywhere All at Once
Qianqian Wang, Yen-Yu Chang, Ruojin Cai, Zhengqi Li, Bharath Hariharan, Aleksander Holynski, Noah Snavely
arxiv
project page
31
704
4K
0
1
35
GART is an explicit, efficient, and expressive non-rigid articulated object representation.
I loved these little dogs when we developed this project!
Our code is now available!
[Code now available!] Gaussian Articulated Template Model (GART) captures humans and animals from monocular videos (ZJU MoCap and PeopleSnapshot in 30s), renders novel poses and views at 150fps, on a laptop, and facilitates text-to-3D generation.
Project:
0
19
112
1
5
35
After EWA volume spaltting, my favorite surfel comes back!! still remember when i first learned about this in 2020, the surfel sota method at that time is DSS from
@toomanyyifans
Cool results with surfels rather than splats.
~2024 feels like ~2004 all over again (the era of point based rendering)
2
24
209
0
1
33
Our
#CVPR2022
work has a similar insight. More at: The code is public now:
A powerful idea in math (that nobody teaches you directly…):
If you don't know how to map between two "things," you can often map each of them to the same "canonical thing."
Then you can just go from the 1st thing to the canonical thing, and back to the 2nd thing. [1/n]
18
141
752
1
3
25
Very similar to a preprocessing pipeline of our bigger system that will be released next month, stay tuned
Introducing “FlowMap”, the first self-supervised, differentiable structure-from-motion method that is competitive with conventional SfM like Colmap!
IMO this solves a major missing piece for internet-scale training of 3D Deep Learning methods.
1/n
11
103
623
0
2
25
Interesting paper! 🤓
NeRDi: Single-View NeRF Synthesis with Language-Guided Diffusion as General Image Priors
abs:
2
44
186
0
0
19
Tomorrow, the first equivariant vision workshop! including both the theory and the practice.
Our Equivariant Vision workshop features five great speakers
@erikjbekkers
@HaggaiMaron
@ninamiolane
@_machc
, and Leo Guibas, spotlight talks, posters, and a tutorial prepared for the vision audience. Come tomorrow, Tuesday, at 8:30am in Summit 321! Thank you
@CongyueD
for
0
15
66
1
1
17
How to generate URDFs of diverse articulated objects? We are going to present this afternoon. Welcome to our poster!
Today
@JiahuiLei1998
and
@CongyueD
will present our Neural Articulation Prior.
@NeurIPSConf
Hall B1+B2 poster
#2015
:
0
1
7
0
2
12
This defines what is a good and impressive work
Code dropped: "Level Of Gaussians
Real-Time View Synthesis for Large Scenes with Millions of Square Meters"
Project:
Code:
9
125
667
0
2
13
This is still my favorite paper today!
CVPR 2015 Best Paper Award
DynamicFusion: Reconstruction and Tracking of Non-rigid Scenes in Real-Time
@rapideRobot
, Dieter Fox, Steve M. Seitz
1
8
77
1
0
10
Introducing our
#CVPR2022
work! We reconstruct deformable geometries (4D) via unsupervised canonicalization -- the factorization of the deformation with the invertible canonical maps and the canonical shape. Many thanks to my advisor
@KostasPenn
Our
#CVPR2022
paper introduces a representation of canonical shape for deformed humans, animals and articulated objects via neural homeomorphisms. Very proud of first paper with my student
@Jiahui77036479
:
arxiv:
Project:
4
16
118
0
0
7
Welcome to our poster on Tue!
Will be presenting our spotlight paper Banana🍌 at
#NeurIPS2023
, welcome to drop by our poster!
Time: Dec 12 (Tue) afternoon
Loc: Poster
#2021
Web:
Presenting one of my favorite works on my birthday🎂sounds like a special experience
3
13
116
0
1
4
Very interesting work
Dexterous manipulation transfer wishes to transfer human manipulations to dexterous robot hand simulations and is inherently difficult due to its intricate, highly-constrained, and discontinuous dynamics and the need to control a dexterous hand with a DoF to accurately replicate
0
1
9
0
0
4
This learned single object equivariant shape prior can be directly used to inference on real scenes for object segmentation with our simple EM algorithm!
0
0
3
This is really interesting
PhyRecon: Physically Plausible Neural Scene Reconstruction
Junfeng Ni, Yixin Chen, Bohan Jing, Nan Jiang, Bin Wang, Bo Dai, Yixin Zhu, Song-Chun Zhu, Siyuan Huang
tl;dr: SP-MC->transformation between SDF-based implicit surface&explicit surface points
2
7
66
0
0
2
@EpisodeYang
This paper is a real “CVPR best paper quality” paper in my mind for long time. The only sad thing about it is for ~10 years, the community has no open-sourced re-implementation for it, all are closed. The nearest open source alternative is SurfelWarp from MIT at RSS2018
0
0
2
@drsrinathsridha
@BrownUniversity
@BrownVisualComp
@BrownBigAI
@BrownHCRI
@BrownCSDept
Congratulations!
0
0
1
Learn SIM(3) equivariant shape prior on ShapeNet, the output mesh strictly follows the input point cloud rotation, translation, and scale!
0
1
1
@mauro_ai
@CongyueD
@GuibasLeonidas
@KostasPenn
Thanks for your interest! Yes, because the shape prior we learn has an Encoder-Decoder architecture, there is no need for optimizing the shape latent during inference like an AutoDecoder (DeepSDF), and the inference is purely feed-forward
0
0
1
Happy to see the follow-up work of our combining the rendering with CaDeX
Neural Surface Reconstruction of Dynamic Scenes with Monocular RGB-D Camera
abs:
project page:
TL;DR: a template-free method to recover high-fidelity geometry, motions and appearance of a dynamic scene from a monocular RGB-D camera
3
65
300
0
0
1
@Ghpkishore
Yes, the code is MIT licensed, and hopefully it may be released in July or August, thank you.
0
0
1
very interesting work
Imagine bringing a personal robot home in the future. How fast can you teach it to do new tasks? We introduce ORION — now a robot can imitate manipulation from a single video demonstration!
5
50
268
0
0
1
@drsrinathsridha
@siggraph
This looks quite interesting, I've been thinking about this problem for a while but never got to a practical solution.
1
0
1