Yushi LAN
@GROS17121524
Followers
369
Following
994
Media
12
Statuses
82
Ph.D. in CV & Graphics
VGG@Oxford, UK
Joined November 2016
Feel free to try STream3R on our github:
github.com
Contribute to NIRVANALAN/STream3R development by creating an account on GitHub.
0
0
5
Our method achieves new SoTA on video depth estimation, and supports window-based inference in real time (with constant VRAM cost).
0
0
4
Specifically, STream3R introduces an streaming framework that processes image sequences efficiently using causal attention, inspired by advances in modern LLMs.
0
0
5
๐ฅStreaming-based 3D/4D Foundation Model๐ฅ. We present STream3R, which reformulates dense 3D/4D reconstruction into a sequential registration task with **causal attention**. - Projects: - Code: - Model:
huggingface.co
8
16
98
RT @zeqi_xiao: While recent works like Genie 2, The Matrix, and Navigation World Models explore video generative models as world simulatorsโฆ.
0
9
0
RT @TheYihangLuo: ๐ฅ Consistent Multi-View Diffusion for 3D Enhancement ๐ฅ. Introducing our work #3DEnhancer @CVPR: a multi-view diffusion moโฆ.
0
10
0
RT @Frozen_Burning: 3DTopia-XL for PBR asset generation is accepted to #CVPR2025๐. The training code has been open-sourced including our **โฆ.
0
53
0
RT @ZhouXia1212: ๐ Introducing ObjCtrl-2.5D: a training-free object motion control model! ๐. Allows more accurate and versatile object motiโฆ.
0
3
0
Our gradio demo generates 2DGS with TSDF mesh from single-view image.
0
0
1
Thanks Xingang! . We have open source the code. Check it below!. -Code: - Demo ๐ค: - Paper:
arxiv.org
While 3D content generation has advanced significantly, existing methods still face challenges with input formats, latent space design, and output representations. This paper introduces a novel 3D...
Introducing ๐๐๐ฎ๐ฌ๐ฌ๐ข๐๐ง ๐๐ง๐ฒ๐ญ๐ก๐ข๐ง๐ , a new 3D generative model with two key properties:.- A structured point-cloud latent space enabling flexible editing!.- Support multi-modal conditions, e.g., point cloud, text, single/multi-view images. arXiv:
3
22
114
RT @XingangP: Introducing ๐๐๐ฎ๐ฌ๐ฌ๐ข๐๐ง ๐๐ง๐ฒ๐ญ๐ก๐ข๐ง๐ , a new 3D generative model with two key properties:.- A structured point-cloud latent space enaโฆ.
0
50
0
RT @XingangP: Introducing ๐๐๐๐๐, which tokenizes 3D objects into multiscale tokens and generates 3D objects by autoregressive next-scale prโฆ.
0
52
0
@CVPR I have release the code here: Feel free to try the colab demo!.
github.com
[CVPR 2023] E3DGE: An encoder-based 3D StyleGAN inversion framework. - NIRVANALAN/CVPR23-E3DGE
0
1
3
Want to dive more into fast๏ผgeometry aware inversion and semantic editing of 3D GANS? Let's talk about it this afternoon @CVPR , poster 032!
1
3
17
RT @jon_barron: #CVPR2023 is mostly posters this year, which is good! Poster sessions can be great. But a lot of people are not so great atโฆ.
0
94
0
RT @_akhaliq: EVA3D: Compositional 3D Human Generation from 2D Image Collections. abs: .project page: https://t.co/โฆ.
0
98
0