Zhiwen(Aaron) Fan @WayneINR Twitter profile

Pinned Tweet

Zhiwen(Aaron) Fan

2 months

🚀 Just dropped the code: #InstantSplat ! Reconstruct your 3D scenes in seconds. Get it now at and start building! ✨ #3DVision #AI #ComputerVision #NVlabs #OpenSource

8

73

362

Last Seen Profiles

@DownVigo

@stwmaniax

@LaVaranda

@BedTorri

@hmonte68negro

@cpsheuz

@Hijabbacol2883

@sayerDrilling30

@ThankTheMaker

@Shadow_Gent1e

@Jan188792Jan

@aalzyadie

@ddu_bbh

@earthtobeca

@GachambiNderit2

@anetaniezajac

@CEREN_Helsinki

@BinorRaja

@ela_ramer

@camilasaTV

@SynthesisAudio

@stw_pdg

@Husnainrafique_

@podcastplaylist

@lorelaigilwore

@TonScalia

@RLCscienceboss

@MarwaElsayedAh1

@mehman1403

@EhDeUwU

@qimingtaimafanl

@_shishi_atiny

@Ddmasd

@yahyacepni

@Pasutriflores12

@RKilipe29309

Zhiwen(Aaron) Fan

@WayneINR

10 months

Excited to present our new paper "LightGaussian": A Compact & Efficient Pipeline for Converting #3D #GaussianSplatting into a more compact format! ✍️15x compression rate ✍️200+FPS Project page 👉: Paper 📷:

4

97

436

Zhiwen(Aaron) Fan

@WayneINR

6 months

Speeding your view synthesis(<40s) with #InstantSplat ! Our large-scale, pose-free method trains in just 37 seconds from sparse views—no #COLMAP , no intrinsics needed. Achieving nearly 30dB test PSNR with just 12 images, New standard in #NVS and new training efficiency.

6

34

207

Zhiwen(Aaron) Fan

@WayneINR

6 months

Our feature3DGS is selected as #CVPR2024 highlight🥳 See you in Seattle. Unlock the answer of 3D Gaussian Splatting + Large 2D Vision Models? - distills feature from 2D large-scale models, to 3D, enabling semantic, editable, and promptable explicit (of course real-time) 3D

2

21

174

Zhiwen(Aaron) Fan

@WayneINR

10 months

Construct your 3D model from just three views? See our "FSGS": A Few-Shot View Synthesis Framework based on #3D #GaussianSplatting ✍️: 2,000x faster than NeRFs ✍️: SSIM from 0.582 (3D-GS) to 0.682 Project page📸: Paper📸:

2

26

133

Zhiwen(Aaron) Fan

@WayneINR

6 months

Exciting update: Our latest results using THREE training views! Despite unknown poses and intrinsics, #instantsplat achieve this in under 20 seconds— even faster under sparser views.

Zhiwen(Aaron) Fan

@WayneINR

6 months

Speeding your view synthesis(<40s) with #InstantSplat ! Our large-scale, pose-free method trains in just 37 seconds from sparse views—no #COLMAP , no intrinsics needed. Achieving nearly 30dB test PSNR with just 12 images, New standard in #NVS and new training efficiency.

6

34

207

10

12

135

Zhiwen(Aaron) Fan

@WayneINR

6 months

Excited to share our #CVPR2024 paper 'Lift3D'. Discover how to elevate 2D vision models to produce #3D consistent predictions and eliminate flickering post-editing—all in a zero-shot manner. Key Highlights: ✍️Applies to #DINO , #CLIP , style transfer, SR, open vocabulary

3

14

115

Zhiwen(Aaron) Fan

@WayneINR

7 months

Totally feeling this #DiT moment with #SORA in 3D Computer Vision! 🚀 The #dust3r , super simple and a game-changer. It flips the usual 3D reconstruction steps on their head, letting us get dense reconstructions in just one go. See video below and more⬇️

14

17

113

Zhiwen(Aaron) Fan

@WayneINR

3 months

Heading to ✈️ #CVPR2024 in #Seattle . Please DM me if you’d like to have a ☕️ chat, especially on efficient 3D modeling, and 3D and 4D generation. Check out the video to see our CVPR presentations and recent work on “Reconstructing Semantic 3D from Unposed Images in Milliseconds”

1

7

113

Zhiwen(Aaron) Fan

@WayneINR

8 months

#GaussianSplatting meets #UTTower 🏰🚁 From ~280 drone-captured images, we've reconstructed an impressive #3DGS model. Our #LightGaussian compresses the model from 398MB to just 26MB! Incredible collaboration with @aaronbuildsmeta @KevinWang_111 🛠️ #3DModeling #TechInnovation

5

16

101

Zhiwen(Aaron) Fan

@WayneINR

3 months

Three of my papers have been accepted by #ECCV2024 (and another one for #IROS )! While paper count isn’t everything, these works explore fascinating new areas: 3D reconstruction from sparse views, high-quality 3D multi-task learning, and large-scale 3D generation using

4

12

88

Zhiwen(Aaron) Fan

@WayneINR

6 months

Another in-the-wild #InstantSplat test using just THREE training views, under resolution of 1920x1080. This demonstrates that proper initialization and disabling Adaptive Density Control effectively suppress excessive 3D Gaussians. We're close to supporting arbitrary

Zhiwen(Aaron) Fan

@WayneINR

6 months

Speeding your view synthesis(<40s) with #InstantSplat ! Our large-scale, pose-free method trains in just 37 seconds from sparse views—no #COLMAP , no intrinsics needed. Achieving nearly 30dB test PSNR with just 12 images, New standard in #NVS and new training efficiency.

6

34

207

4

15

88

Zhiwen(Aaron) Fan

@WayneINR

6 months

- Dive into our latest research on #Multimodal #SLAM : unposed camera images + inertial measurements -> real-time rendering + dense 3D map. - The new UT-MM dataset, a multi-modal one from a mobile robot + a camera + an inertial measurement unit. 👉Paper:

2

15

80

Zhiwen(Aaron) Fan

@WayneINR

6 months

- Three training views on #DL3DV -10K datasets - Camera poses & intrinsics are unknown - Rendering by interpolation - Resolution: 1920x1080 I can tell, pseudo-views are needed to enhance the quality.

Zhiwen(Aaron) Fan

@WayneINR

6 months

Speeding your view synthesis(<40s) with #InstantSplat ! Our large-scale, pose-free method trains in just 37 seconds from sparse views—no #COLMAP , no intrinsics needed. Achieving nearly 30dB test PSNR with just 12 images, New standard in #NVS and new training efficiency.

6

34

207

3

12

79

Zhiwen(Aaron) Fan

@WayneINR

7 months

Just learned #Sora refines prompts to boost user input. Similarly, our latest study, co-led with @Ir1dXD and @ShijieZhoucla , introduces prompt self-refinement along with panoramic #GaussianSplatting for 360 #text -to-3D generation.

6

12

71

Zhiwen(Aaron) Fan

@WayneINR

8 months

The LightGaussian GitHub code has been finalized! If your goal is better model accuracy preservation, selecting a lower compression ratio is advisable. It can still lead to a ~87% reduction in model size and improve FPS from 192 to 244, than the vanilla 3D #GaussianSplatting .

Zhiwen(Aaron) Fan

@WayneINR

10 months

Excited to present our new paper "LightGaussian": A Compact & Efficient Pipeline for Converting #3D #GaussianSplatting into a more compact format! ✍️15x compression rate ✍️200+FPS Project page 👉: Paper 📷:

4

97

436

1

9

56

Zhiwen(Aaron) Fan

@WayneINR

2 months

Four years ago at Alibaba Group, I was part of the team that developed one of the first large-scale view synthesis products, called Holographic World (全息世界). This product had a significant impact during COVID-19 by helping to prevent cross-infection during visits and tours.

MrNeRF

@janusch_patas

2 months

Here is a great example from @felixkit , who created an interactive art gallery with 3DGS. You can walk through the exhibition and get information about the art. I think this sells better than images. The links to the writeup and to the interactive exhibition are below ⬇️

5

33

215

2

5

53

Zhiwen(Aaron) Fan

@WayneINR

7 months

Thrilled to announce that three of our papers have been accepted by #CVPR2024 : Feauture 3DGS, Entropic Score Distillation, and Lift3D. However, I'm even prouder that two of our papers on Efficient 3D Gaussian Splatting were rejected by CVPR2024: #LightGaussian , #FSGS .

8

1

51

Zhiwen(Aaron) Fan

@WayneINR

6 months

The rendering traj is diverging🆘. - THREE training views (by random) on a custom scene. - Just three images, no assumption on poses, camera parameters. - IPhone, 1428x1071 resolution But controlling the rendering traj to be reasonable is really a little bit hard.

Zhiwen(Aaron) Fan

@WayneINR

6 months

Speeding your view synthesis(<40s) with #InstantSplat ! Our large-scale, pose-free method trains in just 37 seconds from sparse views—no #COLMAP , no intrinsics needed. Achieving nearly 30dB test PSNR with just 12 images, New standard in #NVS and new training efficiency.

6

34

207

4

5

47

Zhiwen(Aaron) Fan

@WayneINR

6 months

(2/N) 🌐 New paradigm of 3D #AIGC !📷 A simple text prompt "A well-known landmark in Italy", and here is your generated 3D panoramic #GaussianSplatting

2

3

44

Zhiwen(Aaron) Fan

@WayneINR

2 months

Amazing slides for anyone wanting to learn GS from scratch. Thanks, Forrest! Special thanks for highlighting our #LightGaussian work, where visibility-based importance scoring to prune 'least-important' Gaussians is crucial for rendering efficiency.

Forrest Iandola

@fiandola

2 months

Here's my CVPR tutorial on 𝗚𝗮𝘂𝘀𝘀𝗶𝗮𝗻 𝗦𝗽𝗹𝗮𝘁𝘁𝗶𝗻𝗴. Enjoy!

5

82

310

2

39

Zhiwen(Aaron) Fan

@WayneINR

3 months

Thanks for featuring our work! #4K4DGen supports exporting 4K video in RGB+depth, easily integrating into VR/MR headsets.

MrNeRF

@janusch_patas

3 months

4K4DGen: Panoramic 4D Generation at 4K Resolution Project:

6

27

127

1

4

34

Zhiwen(Aaron) Fan

@WayneINR

6 months

Elevate your #Sora experience to another dimension! 🌟 Dive into a 3D world with our virtual art gallery, where panorama images transform into 360-degree 3D Gaussian Splatting wonders 🎨✨ Explore art like never before - in real-time! #3DWorld #ArtExploration #InnovateWithUs

3

4

31

Zhiwen(Aaron) Fan

@WayneINR

1 month

Interesting and smart application. Accurate counting can be done only in 3D space for fruit trees and something similar…

MrNeRF

@janusch_patas

1 month

FruitNeRF: A Unified Neural Radiance Field based Fruit Counting Framework Project: Code: Uses NeRFs to count any fruit type directly in 3D. The code is a Nerfstudio extension! Method ⬇️ 1 | 2

11

116

626

0

3

27

Zhiwen(Aaron) Fan

@WayneINR

8 months

First day at Autonomous Vehicle Research @NVIDIA , in collaboration with @yuewang314 and @iamborisi . Eager to deliver impactful advancements in efficient #3D modeling/generation. 🤗🤗

1

0

25

Zhiwen(Aaron) Fan

@WayneINR

8 months

Exploring the capabilities of #DepthAnything – it’s impressive how it enhances drone-view videos (OOD I guess) by delivering reliable and accurate monocular depth values. What else can this technology offer?

0

24

Zhiwen(Aaron) Fan

@WayneINR

5 months

To consider the optimization "freedom": the better initialization -> the easier optimization -> the faster training speed -> the detailed rendering quality

MrNeRF

@janusch_patas

5 months

Does Gaussian Splatting need SFM Initialization?

0

18

103

2

22

Zhiwen(Aaron) Fan

@WayneINR

5 months

Great progress in enabling end-to-end, differentiable 3D vision! It's a privilege to witness and contribute to these exciting developments in the field.

Vincent Sitzmann

@vincesitzmann

5 months

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

102

627

1

19

Zhiwen(Aaron) Fan

@WayneINR

2 months

Fortunately, today we have advanced tools like NeRF and 3DGS that allow us to efficiently capture posed RGB-D inputs and convert them into Holographic World without the need for a long and tedious pipeline. By leveraging large-scale pretrained models (DUSt3R, Monocular Depth,

0

1

19

Zhiwen(Aaron) Fan

@WayneINR

6 months

Thank you, AK, for highlighting our team's recent advancements in optimizing the view synthesis pipeline!🤗 Visit the project webpage:

AK

@_akhaliq

6 months

InstantSplat Unbounded Sparse-view Pose-free Gaussian Splatting in 40 Seconds While novel view synthesis (NVS) has made substantial progress in 3D computer vision, it typically requires an initial estimation of camera intrinsics and extrinsics from dense viewpoints. This

3

23

177

2

0

18

Zhiwen(Aaron) Fan

@WayneINR

6 months

@LuLing26466911 Late April or early May! Going through the releasing process, as well as adding more features on it!

1

0

12

Zhiwen(Aaron) Fan

@WayneINR

6 months

@janusch_patas @JeromeRevaud Check our paper:

InstantSplat: Sparse-view SfM-free Gaussian Splatting in Seconds

While novel view synthesis (NVS) from a sparse set of images has advanced significantly in 3D computer vision, it relies on precise initial estimation of camera parameters using...

arxiv.org

0

12

Zhiwen(Aaron) Fan

@WayneINR

6 months

@janusch_patas @JeromeRevaud Hi, thanks for posting! Our paper will be online very soon. tl:dr: proper pre-processing DUSt3R + 3DGS -> 37 seconds for pose-free, intrinsics-free, sparse-view novel-view synthesis, with ~30dB PSNR.

1

0

12

Zhiwen(Aaron) Fan

@WayneINR

5 months

Explore immersive Mars footage with the #Perseverance rover from @NASA . We can recover the 3D from #Perseverance 's stereo navigation cameras! Detailed 3D reconstructions can be done in just ~20 seconds from scratch, without using the calibration from @NASA_Technology .

1

3

11

Zhiwen(Aaron) Fan

@WayneINR

5 months

@xiaolonw @OpenAI Continuously inspired! Following the impressive in-the-wild test cases, we've experimented with InstantSplat (), training it on 12 views from the Sora video without pre-computing camera parameters from COLMAP.

0

1

11

Zhiwen(Aaron) Fan

@WayneINR

4 months

Finally, #4D #generation at #4K resolution. But X resizes it back to 1080p🤣

0

1

12

Zhiwen(Aaron) Fan

@WayneINR

2 months

Was wondering the system pipeline from images to 3D GS/meshes, is it still COLMAP+GS?

Jan Verwoerd

@jan_bsc_

2 months

Ah yes! Finally we’ve finished the Gaussian Splatting - Turn table capture workflow. No more walking around the object or get a huge mess of clouds when doing Gaussian Splats on our Marc scanner. This is a huge deal for us because it will enable us to scan, process, reconstruct

6

13

158

0

10

Zhiwen(Aaron) Fan

@WayneINR

6 months

Interesting, densification is the core factor for the final rendering quality of 3DGS, but only few papers pay attention to

MrNeRF

@janusch_patas

6 months

Revising Densification in Gaussian Splatting

0

5

82

0

3

10

Zhiwen(Aaron) Fan

@WayneINR

5 months

AI patents number is important but is not everything... We see a lot "Breakthroughs" researches from European teams in recent years, including but not limited to "3D Gaussian Splatting", "DUSt3R"...

艾森 Essen

@essen_ai

5 months

不同国家和地区每年授予的AI专利数量对比：AI时代中美竞跑，谁最终胜出🏆现在还不好说，但有一点已经可以非常肯定，那就是：没欧洲什么事了。老欧洲现在只能靠全球最早推出AI监管立法来找存在感了...

31

8

39

1

0

9

Zhiwen(Aaron) Fan

@WayneINR

5 months

0

9

Zhiwen(Aaron) Fan

@WayneINR

5 months

Travel to DC, just learned a little more about #Sora . Below is the details ⬇️

3

0

8

Zhiwen(Aaron) Fan

@WayneINR

6 months

🤣

Julian Bilcke

@flngr

6 months

3

7

61

1

0

8

Zhiwen(Aaron) Fan

@WayneINR

5 months

@Snosixtytwo Congratulations! Any plan in code releasing?

1

0

7

Zhiwen(Aaron) Fan

@WayneINR

7 months

Sora -> 4D Modeling -> Free exploration. A lot of possibilities.

Radiance Fields

@RadianceFields

7 months

Sora unlocks generative large scale radiance fields and this is such a massive deal

9

61

317

0

1

8

Zhiwen(Aaron) Fan

@WayneINR

10 months

🧐Ensure your #3D #GaussianSplatting is compatible with various downstream tasks. 📸Explore our #3DGaussian #FeatureField that is distilled from 2D large-scale models! 📸Semantic scene understanding, language-guided editing, and 3D #SegAnything are now being explored.

Shijie Zhou

@ShijieZhoucla

10 months

(1/5) 3D Gaussian Splatting + AI Foundation Models = ? 🤔 Feature 3DGS🪄, distills feature fields from 2D foundation models, opening the door to a brand new semantic, editable, and promptable explicit 3D scene representation. AI for 2D, now in 3D! 🚀 🔗

1

22

84

0

8

Zhiwen(Aaron) Fan

@WayneINR

4 months

wow

MrNeRF

@janusch_patas

4 months

Tetrahedron Splatting for 3D Generation

2

32

156

0

7

Zhiwen(Aaron) Fan

@WayneINR

8 months

Happy Lunar New Year! 🎉 May the year ahead bring prosperity, joy, and good news, just like this inspiring image from a WeChat moment. Wishing everyone a year filled with happiness and success. #LunarNewYear #GoodVibes

5

0

7

Zhiwen(Aaron) Fan

@WayneINR

7 months

Check our paper published on #3DV that utilizes the #DINO ViT feature multiple stages classification for few-shot #6DoF object pose estimation: ✍️ Cas6D: Learning to Estimate 6DoF Pose from Limited Data: A Few-Shot, Generalizable Approach using RGB Images

International Conference on 3D Vision

@3DVconf

7 months

The list of accepted papers and abstract is now out! Interested in catching up with the state-of-the-art 3D vision? Check it out!!

0

12

46

0

6

Zhiwen(Aaron) Fan

@WayneINR

6 months

(1/N) 🌐 Take your #Sora journey to new heights! 🚀 Unveil a universe where 3D panoramic #GaussianSplatting meet creativity - think "colossal, man-shaped cloud towering over the earth"! 🌩️ Dive into a real-time virtual realm and let your imagination soar! 🎨 #3DExperience

0

6

Zhiwen(Aaron) Fan

@WayneINR

1 year

Come see our poster presentation tomorrow (Wednesday, 4th, 10:30 AM - 12:30 PM) at Vision and graphics section!

AK

@_akhaliq

2 years

StegaNeRF: Embedding Invisible Information within Neural Radiance Fields abs: project page:

3

15

88

6

1

6

Zhiwen(Aaron) Fan

@WayneINR

6 months

Discover the magic when #Starship meets #AIGC ! 🚀 With our new text-to-3D tech, words craft #3D realms—envision Starship's #Mars landing soon. @elonmusk , how about integrating a real Starship model to the virtual world? 😄

0

1

6

Zhiwen(Aaron) Fan

@WayneINR

6 months

@janusch_patas Thanks for sharing🥳! We're thinking the issue of the optimal representation for large-scale scene generation. The ideal representation should manage unbounded scenarios without content duplication—a common issue with outpainting due to its lack of global awareness.

0

6

Zhiwen(Aaron) Fan

@WayneINR

7 months

(5/N). Few-shot Gaussian Splatting () was led by me, ZehaoZhu and @YifanJiang17 This method achieves the best sparse-view NVS performance, and it runs over 2000 times faster than NeRF-based methods.

0

6

Zhiwen(Aaron) Fan

@WayneINR

5 months

This is BIG and insightful

Jerome Revaud

@JeromeRevaud

5 months

@WayneINR @AjdDavison @LourdesAgapito Here are the slides of my recent invited talk about CroCo + DUSt3R + MASt3R Note: It's a pdf, i know the videos won't work but there's really nothing hitherto unseen. I can point to where to find each individual video upon request.

5

10

85

0

6

Zhiwen(Aaron) Fan

@WayneINR

7 months

Excited to explore how this breakthrough in sparse view synthesis will revolutionize the field! I’ll experiment with it and report the findings soon.

Jerome Revaud

@JeromeRevaud

7 months

An example of how DUSt3R can do "impossible matching": given two images without any shared visual content (my office, obviously never seen at training), it can output an accurate reconstruction (no intrinsics, no poses!) in seconds

18

107

457

0

6

Zhiwen(Aaron) Fan

@WayneINR

6 months

@PMel3D Working on this🫡

1

0

4

Zhiwen(Aaron) Fan

@WayneINR

6 months

Worked together with @sneezygiraffe @peihao_wang , among others, on our recent project. See you in #Seattle 🤗

0

3

Zhiwen(Aaron) Fan

@WayneINR

7 months

(1/N). Feauture 3DGS: Supercharging 3D Gaussian Splatting to Enable Distilled Feature Fields, where it enables the distillation of 3D feature fields from any 2D foundation models, using only 18 training views for this complex real-world scene. Website:

0

5

Zhiwen(Aaron) Fan

@WayneINR

1 year

@_akhaliq

0

3

Zhiwen(Aaron) Fan

@WayneINR

10 months

Hello #NewOrleans again 🎉 see you all in #NeurIPS 2023.

0

4

Zhiwen(Aaron) Fan

@WayneINR

7 months

(7/N). Ackowledgement: Both REJECTED papers are open-sourced, and thank to the reviewers for the valuable suggestions, and thanks to Hugging Face to invite us to deploy models on their platforms. We will continue to revise our drafts and address all the reviewers' comments.

0

4

Zhiwen(Aaron) Fan

@WayneINR

7 months

No more Image Correspondence -> Sparse Reconstruction -> Dense Reconstruction. Just one feedforwad! And it's crushing it with OOD data – just 12 drone pics and you've got your model. Mind blown! 🤯

0

4

Zhiwen(Aaron) Fan

@WayneINR

7 months

(4/N). The two rejected papers are even more interesting: LightGaussian () was led by me, @KevinWang_111 , and @KairunWen . Our method ensures rendering quality while achieving 15x model compression and 50% rendering efficiency, than #GaussianSplatting

0

4

Zhiwen(Aaron) Fan

@WayneINR

7 months

Seeking PMs' insights on the LPIPS curve's alignment with human visual perception. The figures illustrate how LightGaussian maintains rendering quality even when up to 60% of Gaussians are pruned. A breakthrough in #GaussianSplatting . See the visual comparison below.

Zhiwen(Aaron) Fan

@WayneINR

8 months

The LightGaussian GitHub code has been finalized! If your goal is better model accuracy preservation, selecting a lower compression ratio is advisable. It can still lead to a ~87% reduction in model size and improve FPS from 192 to 244, than the vanilla 3D #GaussianSplatting .

1

9

56

1

0

3

Zhiwen(Aaron) Fan

@WayneINR

7 months

(3/N). Lift3D: Zero-Shot Lifting of Any 2D Vision Model to 3D, where we extend a single 2D vision operator into consistent 3D under a generalizable way.

0

4

Zhiwen(Aaron) Fan

@WayneINR

7 months

I cannot imagine this is generated one wo reading the description. How Sora controls the camera motion BTW?

OpenAI

@OpenAI

7 months

Introducing Sora, our text-to-video model. Sora can create videos of up to 60 seconds featuring highly detailed scenes, complex camera motion, and multiple characters with vibrant emotions. Prompt: “Beautiful, snowy

10K

32K

138K

0

4

Zhiwen(Aaron) Fan

@WayneINR

6 months

Datasets credit: DL3DV-10K. Scene ID: ba55c875d20c34ee85ffc72264c4d77710852e5fb7d9ce4b9c26a8442850e98f

0

1

4

Zhiwen(Aaron) Fan

@WayneINR

7 months

R3: Although promising video results have been achieved, what’s the difference between yours and the previous image generation one?

Jia-Bin Huang

@jbhuang0604

7 months

R2: While the results are impressive, this is a simple combination of diffusion transformer (ICCV 2023) and latent diffusion model (CVPR 2022). Limited novelty. Weak reject.

55

155

2K

0

4

Zhiwen(Aaron) Fan

@WayneINR

10 months

UTers with Stanford friend@ #NeurIPS2023

0

1

4

Zhiwen(Aaron) Fan

@WayneINR

6 months

@JeromeRevaud @janusch_patas @PMel3D That's exciting direction! We're also advancing towards making 3D modeling as seamless and efficient as its 2D counterpart, enabling end-to-end capabilities.

1

0

4

Zhiwen(Aaron) Fan

@WayneINR

5 months

Side parking🫣

Massimo

@Rainmaker1973

5 months

Different techniques of side parking. [🎞️ carknowledge01]

413

6K

29K

1

0

3

Zhiwen(Aaron) Fan

@WayneINR

10 months

See you in NeurIPS

VITA Group

@VITAGroupUT

10 months

#NeurIPS2023 here we go again, New Orleans!! We're excited to meet you and to present our latest findings in #LLM , #diffusion , #mlsys , #graph , and more. Please check out our #NeurIPS activity at-a-glance. w/ @KyriectionZhang @YifanJiang17 @WayneINR @TianlongChen4 @ShiweiLiu9

1

4

17

0

3

Zhiwen(Aaron) Fan

@WayneINR

2 months

通过能产品化的Neuralink改变交互方式，能颠覆现今所有的教育行业，沟通能力甚至人类进化速度……

阑夕

@foxshuo

2 months

马斯克带着脑机接口公司Neuralink的核心团队，以及植入人脑芯片的第一个受试者，参加了MIT科学家Lex Fridman的播客，一聊就是8个多小时，是的，你没看错，全程8个小时的对谈，信息密度极高…… 可以看得出来，马斯克很欣赏Lex Fridman，这是他第5次做客Lex

53

413

1K

0

3

Zhiwen(Aaron) Fan

@WayneINR

6 months

All contents are generated from a text prompt, with intermediate representations include panoramic image, panoramic 3D Gaussian Splatting.

0

3

Zhiwen(Aaron) Fan

@WayneINR

7 months

@Ir1dXD @ShijieZhoucla We've achieved seamless scene content creation directly from very simple text. Preprint and more video coming soon!

0

3

Zhiwen(Aaron) Fan

@WayneINR

6 months

@JeromeRevaud @janusch_patas @PMel3D cf-3dgs is internally open-sourced at NVIDIA..

1

0

2

Zhiwen(Aaron) Fan

@WayneINR

2 months

Come and join the venue in NeurIPS!

Junyuan "Jason" Hong

@hjy836

2 months

📢Announcing the first GenAI4Health Workshop at #NeurIPS2024 where we invite speakers and participants from #health , #AI_safety , and #AI_policy areas to discuss the entangled challenges of potential, trust, and policy compliance of GenAI4Health! 🌍Homepage:

1

14

36

0

3

Zhiwen(Aaron) Fan

@WayneINR

5 months

top-down view of the 3D scene representation

0

3

Zhiwen(Aaron) Fan

@WayneINR

7 months

How can we unlock the possibilities from #dust3r for #3dmodeling #GaussianSplatting ?

1

0

3

Zhiwen(Aaron) Fan

@WayneINR

10 months

Cool! I’ll try it with my data and am curious to see how it performs with outdoor datasets.

MrNeRF

@janusch_patas

10 months

Just tried SplaTAM and it's really cool! Took a moment to get the hang of capturing my desk with the iPhone accurately - it definitely requires some skill. But, so does navigating Open3D :). Check out their demo to get a peek at what the future holds!

4

13

91

0

3

Zhiwen(Aaron) Fan

@WayneINR

9 months

@patriciogv How was the same comparisons with metric depth estimator:

GitHub - YvanYin/Metric3D: The repo for "Metric3D: Towards Zero-shot Metric 3D Prediction from A...

The repo for "Metric3D: Towards Zero-shot Metric 3D Prediction from A Single Image" and "Metric3Dv2: A Versatile Monocular Geometric Foundation Model..." - YvanYin/Metric3D

github.com

1

0

3

Zhiwen(Aaron) Fan

@WayneINR

6 months

@janusch_patas Thanks for posting🤗

0

3

Zhiwen(Aaron) Fan

@WayneINR

10 months

LOL really want to design a module to make the moving part looks more consistent

Benjamin De Kraker 🏴‍☠️

@BenjaminDEKR

10 months

Experimenting with Magic-Animate and I'm so, so sorry 😆😆😆

36

10

161

0

3

Zhiwen(Aaron) Fan

@WayneINR

7 months

(2/N). Taming Mode Collapse in Score Distillation for Text-to-3D Generation: we tackle the Janus problem in text-to-3D generation. Website:

0

3

Zhiwen(Aaron) Fan

@WayneINR

10 months

@kwea123 Hi, thanks for sharing your concerns. Regarding the question, I partially (about 20%) agree with your opinion: 1. "The difficulty of comparison of the quality": I understand this point, as multiple teams are working independently, and there's no standard benchmark for

0

3

Zhiwen(Aaron) Fan

@WayneINR

7 months

It’s just the moment of #Sora v1, and it definitely will surprise us with its updated version

Jim Fan

@DrJimFan

7 months

I see some vocal objections: "Sora is not learning physics, it's just manipulating pixels in 2D". I respectfully disagree with this reductionist view. It's similar to saying "GPT-4 doesn't learn coding, it's just sampling strings". Well, what transformers do is just manipulating

249

460

3K

0

3

Zhiwen(Aaron) Fan

@WayneINR

5 months

@JeromeRevaud @AjdDavison @LourdesAgapito When will we see more details about MASt3R🫣

1

0

2

Zhiwen(Aaron) Fan

@WayneINR

6 months

@GKopanas @PapantonakisP Cool work! Pruning, SH assignment/VQ are for both primitive redundancy and feature redundancy. Our #LightGaussian shares very similar high-level idea with it.

0

2

Zhiwen(Aaron) Fan

@WayneINR

6 months

@Nik__V__ @JeromeRevaud @yuewang314 Check the frame in 20s (), which is the rendered image from the initialized 3DGS, using DUSt3R.

1

0

2

Zhiwen(Aaron) Fan

@WayneINR

6 months

@JonClark55 @LuLing26466911 We primarily focus on high-resolution, large-scale scenes. I believe adapting our approach to object-level datasets should not be difficult, as demonstrated by our results on the MVImgNet dataset in our paper.

2

0

2

Zhiwen(Aaron) Fan

@WayneINR

1 year

Much of the current research focus is on improving the quality of nerf. However, there is one important research question ignored by the current nerf development. The question is can we apply steganography in nerf?🧐

0

1

Zhiwen(Aaron) Fan

@WayneINR

4 months

I like Embody more

gaut

@0xgaut

4 months

The majority of the U.S GDP is attributable to this chair

203

742

20K

0

2

Zhiwen(Aaron) Fan

@WayneINR

6 months

@Coding_Black @JeromeRevaud

Zhiwen(Aaron) Fan

@WayneINR

6 months

- Three training views on #DL3DV -10K datasets - Camera poses & intrinsics are unknown - Rendering by interpolation - Resolution: 1920x1080 I can tell, pseudo-views are needed to enhance the quality.

3

12

79

0

1

2

Zhiwen(Aaron) Fan

@WayneINR

6 months

@JeromeRevaud I believe your work will earn a 'test of time' award.

0

2