正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

📢 3D world models from video diffusion suffer from inconsistent frames -> blurry output. Our fix: instead of naïve 3D reconstruction, we non-rigidly align each frame into a globally-consistent 3DGS representation. ->sharp visuals on top of any VDM!

Matthias Niessner

50,221 subscribers

53,323 次观看 • 7 个月前 •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

5 条评论

Matthias Niessner 的头像

Matthias Niessner7 个月前

The core idea of our method is a tailored non-rigid frame-to-model ICP that aligns per-frame pointclouds into a canonical world space. This resolves geometric inconsistencies across frames, which we leverage as initialization for our non-rigid 3D reconstruction. The resulting 3D reconstructions have high visual fidelity without floating artifacts. Our approach is agnostic to many state-of-the-art video models; i.e., we can resolve their inconsistencies and turn them into 3D world generators - check out the results on our project page. Great work by the amazing @LukasHollein

Bohan Zeng 的头像

Bohan Zeng7 个月前

Impressive work! could you please leave a issue in our unified framework: , and we will add you splendid work in our OpenWorldLib

Dmytro Omelian 🇺🇦 的头像

Dmytro Omelian 🇺🇦7 个月前

this feels like the missing glue layer between "pretty good video" and "usable 3d" curious how far 3dgs can stretch before it breaks on long, messy, handheld clips

Hang Chu 的头像

Hang Chu1 个月前

really like the inverse deformation idea. how much does this depend on DA3? in the examples, planar surfaces seem to hold up better than the lamp and sofa in the middle of the room.

puppy 的头像

puppy7 个月前

interesting work!

相关视频

Synthesizing worlds with video diffusion models is often inconsistent — moving the camera back and forth leads to different scenes. We propose 🌐𝗪𝗼𝗿𝗹𝗱𝗠𝗲𝗺, a memory-based approach that ensures consistent world simulation without relying on explicit 3D reconstruction.

Synthesizing worlds with video diffusion models is often inconsistent — moving the camera back and forth leads to different scenes. We propose 🌐𝗪𝗼𝗿𝗹𝗱𝗠𝗲𝗺, a memory-based approach that ensures consistent world simulation without relying on explicit 3D reconstruction.

Xingang Pan

19,413 次观看 • 1 年前

Can we use video diffusion to generate 3D scenes? 𝐖𝐨𝐫𝐥𝐝𝐄𝐱𝐩𝐥𝐨𝐫𝐞𝐫 (#SIGGRAPHAsia25) creates fully-navigable scenes via autoregressive video generation. Text input -> 3DGS scene output & interactive rendering! 🌍 📽️

Can we use video diffusion to generate 3D scenes? 𝐖𝐨𝐫𝐥𝐝𝐄𝐱𝐩𝐥𝐨𝐫𝐞𝐫 (#SIGGRAPHAsia25) creates fully-navigable scenes via autoregressive video generation. Text input -> 3DGS scene output & interactive rendering! 🌍 📽️

Matthias Niessner

30,929 次观看 • 1 年前

📢 Lyra: Generative 3D Scene Reconstruction via Video Diffusion Model Self-Distillation Got only one or a few images and wondering if recovering the 3D environment is a reconstruction or generation problem? Why not do it with a generative reconstruction model! We show that a camera-conditioned video diffusion model can be transformed into a generative reconstruction model that directly outputs a high-quality 3D Gaussian Splatting representation through self-distillation, without requiring real-world training data. Check out our results in the video (wait for dynamic scenes in the second half!) : Project Page: Code and Models: Paper:

📢 Lyra: Generative 3D Scene Reconstruction via Video Diffusion Model Self-Distillation Got only one or a few images and wondering if recovering the 3D environment is a reconstruction or generation problem? Why not do it with a generative reconstruction model! We show that a camera-conditioned video diffusion model can be transformed into a generative reconstruction model that directly outputs a high-quality 3D Gaussian Splatting representation through self-distillation, without requiring real-world training data. Check out our results in the video (wait for dynamic scenes in the second half!) : Project Page: Code and Models: Paper:

Sherwin Bahmani

66,918 次观看 • 1 年前

V3D Video Diffusion Models are Effective 3D Generators Automatic 3D generation has recently attracted widespread attention. Recent methods have greatly accelerated the generation speed, but usually produce less-detailed objects due to limited model capacity or 3D data. Motivated by recent advancements in video diffusion models, we introduce V3D, which leverages the world simulation capacity of pre-trained video diffusion models to facilitate 3D generation. To fully unleash the potential of video diffusion to perceive the 3D world, we further introduce geometrical consistency prior and extend the video diffusion model to a multi-view consistent 3D generator. Benefiting from this, the state-of-the-art video diffusion model could be fine-tuned to generate 360degree orbit frames surrounding an object given a single image. With our tailored reconstruction pipelines, we can generate high-quality meshes or 3D Gaussians within 3 minutes. Furthermore, our method can be extended to scene-level novel view synthesis, achieving precise control over the camera path with sparse input views. Extensive experiments demonstrate the superior performance of the proposed approach, especially in terms of generation quality and multi-view consistency

V3D Video Diffusion Models are Effective 3D Generators Automatic 3D generation has recently attracted widespread attention. Recent methods have greatly accelerated the generation speed, but usually produce less-detailed objects due to limited model capacity or 3D data. Motivated by recent advancements in video diffusion models, we introduce V3D, which leverages the world simulation capacity of pre-trained video diffusion models to facilitate 3D generation. To fully unleash the potential of video diffusion to perceive the 3D world, we further introduce geometrical consistency prior and extend the video diffusion model to a multi-view consistent 3D generator. Benefiting from this, the state-of-the-art video diffusion model could be fine-tuned to generate 360degree orbit frames surrounding an object given a single image. With our tailored reconstruction pipelines, we can generate high-quality meshes or 3D Gaussians within 3 minutes. Furthermore, our method can be extended to scene-level novel view synthesis, achieving precise control over the camera path with sparse input views. Extensive experiments demonstrate the superior performance of the proposed approach, especially in terms of generation quality and multi-view consistency

AK

31,997 次观看 • 2 年前

One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape Optimization paper page: Single image 3D reconstruction is an important but challenging task that requires extensive knowledge of our natural world. Many existing methods solve this problem by optimizing a neural radiance field under the guidance of 2D diffusion models but suffer from lengthy optimization time, 3D inconsistency results, and poor geometry. In this work, we propose a novel method that takes a single image of any object as input and generates a full 360-degree 3D textured mesh in a single feed-forward pass. Given a single image, we first use a view-conditioned 2D diffusion model, Zero123, to generate multi-view images for the input view, and then aim to lift them up to 3D space. Since traditional reconstruction methods struggle with inconsistent multi-view predictions, we build our 3D reconstruction module upon an SDF-based generalizable neural surface reconstruction method and propose several critical training strategies to enable the reconstruction of 360-degree meshes. Without costly optimizations, our method reconstructs 3D shapes in significantly less time than existing methods. Moreover, our method favors better geometry, generates more 3D consistent results, and adheres more closely to the input image. We evaluate our approach on both synthetic data and in-the-wild images and demonstrate its superiority in terms of both mesh quality and runtime. In addition, our approach can seamlessly support the text-to-3D task by integrating with off-the-shelf text-to-image diffusion models.

One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape Optimization paper page: Single image 3D reconstruction is an important but challenging task that requires extensive knowledge of our natural world. Many existing methods solve this problem by optimizing a neural radiance field under the guidance of 2D diffusion models but suffer from lengthy optimization time, 3D inconsistency results, and poor geometry. In this work, we propose a novel method that takes a single image of any object as input and generates a full 360-degree 3D textured mesh in a single feed-forward pass. Given a single image, we first use a view-conditioned 2D diffusion model, Zero123, to generate multi-view images for the input view, and then aim to lift them up to 3D space. Since traditional reconstruction methods struggle with inconsistent multi-view predictions, we build our 3D reconstruction module upon an SDF-based generalizable neural surface reconstruction method and propose several critical training strategies to enable the reconstruction of 360-degree meshes. Without costly optimizations, our method reconstructs 3D shapes in significantly less time than existing methods. Moreover, our method favors better geometry, generates more 3D consistent results, and adheres more closely to the input image. We evaluate our approach on both synthetic data and in-the-wild images and demonstrate its superiority in terms of both mesh quality and runtime. In addition, our approach can seamlessly support the text-to-3D task by integrating with off-the-shelf text-to-image diffusion models.

AK

179,061 次观看 • 3 年前

Today we announced RTFM (Real-Time Frame Model) — a world model that generates frames in real time from any camera viewpoint. Unlike standard video models, RTFM understands 3D geometry. You can literally move the camera through the generated world. 🎥👇

Today we announced RTFM (Real-Time Frame Model) — a world model that generates frames in real time from any camera viewpoint. Unlike standard video models, RTFM understands 3D geometry. You can literally move the camera through the generated world. 🎥👇

Keunhong Park

16,112 次观看 • 1 年前

📢WorldAgents: 3D worlds only from 2D image models - without any training! We propose an agentic approach with a Director (VLM) to plan the scene, a Generator (Flux or NanoBanana) for new views, and a Verifier (VLM) for selection / 3D consistency. -> High-fidelity 3D worlds from a single text prompt. What's remarkable: our agents find consistent views from 2D image models to obtain 3D-consistent worlds; this shows that image models contain world priors - agents just need to find them! Great work by Ziya Erkoç Angela Dai

📢WorldAgents: 3D worlds only from 2D image models - without any training! We propose an agentic approach with a Director (VLM) to plan the scene, a Generator (Flux or NanoBanana) for new views, and a Verifier (VLM) for selection / 3D consistency. -> High-fidelity 3D worlds from a single text prompt. What's remarkable: our agents find consistent views from 2D image models to obtain 3D-consistent worlds; this shows that image models contain world priors - agents just need to find them! Great work by Ziya Erkoç Angela Dai

Matthias Niessner

19,093 次观看 • 7 个月前

📽️➡️🏠 Can an image become a consistent 3D world? Video Diffusion Models look stunning, but 3D consistency is still a nightmare… #WorldStereo is different. We gave the AI a "3D brain" using Geometric Memories. The result? ✅ Zero flickering. ✅ Perfect 3D consistency.

📽️➡️🏠 Can an image become a consistent 3D world? Video Diffusion Models look stunning, but 3D consistency is still a nightmare… #WorldStereo is different. We gave the AI a "3D brain" using Geometric Memories. The result? ✅ Zero flickering. ✅ Perfect 3D consistency.

Tengfei Wang

20,067 次观看 • 7 个月前

Check out CAT3D! Image(s)-to-3D in 1 minute! Given any number of real or generated images, CAT3D uses a multi-view diffusion prior to create consistent novel views. These views are used to reconstruct a 3D scene using NeRF/3DGS.

Check out CAT3D! Image(s)-to-3D in 1 minute! Given any number of real or generated images, CAT3D uses a multi-view diffusion prior to create consistent novel views. These views are used to reconstruct a 3D scene using NeRF/3DGS.

Philipp Henzler

12,752 次观看 • 2 年前

📢📢GenRecon: Bridging Generative Priors for Multi-View 3D Scene Reconstruction📢📢 Reconstructing high-fidelity 3D scenes from sparse RGB input is hard. It needs a strong 3D prior! We reformulate multi-view scene reconstruction as conditional 3D generation over overlapping spatial chunks, lifting posed image features into a generative shape prior via 3D conditioning. As an example prior, we build on Trellis2, and train it such that its reconstruction is pixel aligned and matches from all views. GenRecon achieves unprecedented reconstruction quality from any sparse RGB input sequence, even from a phone capture. The reconstruction also includes PBR materials which facilitates relighting and virtual object insertion. Amazing work by Katharina Schmid, Nicolas von Lützow, Jozef, Angela Dai

📢📢GenRecon: Bridging Generative Priors for Multi-View 3D Scene Reconstruction📢📢 Reconstructing high-fidelity 3D scenes from sparse RGB input is hard. It needs a strong 3D prior! We reformulate multi-view scene reconstruction as conditional 3D generation over overlapping spatial chunks, lifting posed image features into a generative shape prior via 3D conditioning. As an example prior, we build on Trellis2, and train it such that its reconstruction is pixel aligned and matches from all views. GenRecon achieves unprecedented reconstruction quality from any sparse RGB input sequence, even from a phone capture. The reconstruction also includes PBR materials which facilitates relighting and virtual object insertion. Amazing work by Katharina Schmid, Nicolas von Lützow, Jozef, Angela Dai

Matthias Niessner

18,447 次观看 • 5 个月前

$📢📢 Video Diffusion Avatars📢📢 Can video diffusion models (VDMs) replace 3D avatars? Video Diffusion Avatars (VDA) creates photo-realistic avatars from a single image, with full control over camera viewpoint, facial expressions and upper-body pose. Unlike prevailing 3D avatar methods, which commonly rely on 3DGS/NeRF, our VDM-rendering produces physically plausible dynamics of hair and earrings, as well as complex light interactions like shadows, reflections and refractions. Crucially, we show that common 3D representaion (e.g. 3DGS, NeRF) are not necessary for photorealistic and view-point controllable avatars. VDA is quite affordable for the research community, since training only requires 34 A100-GPU hours on public multi-view datasets, yet VDA generalizes far beyond them: to paintings, human-like creatures and full 360° orbits. 🌐 ▶️ Great work by Simon Giebenhain, Tobias Kirschstein, Liam Schoneveld, Davide Davoli, Zhe Chen$

📢📢 Video Diffusion Avatars📢📢 Can video diffusion models (VDMs) replace 3D avatars? Video Diffusion Avatars (VDA) creates photo-realistic avatars from a single image, with full control over camera viewpoint, facial expressions and upper-body pose. Unlike prevailing 3D avatar methods, which commonly rely on 3DGS/NeRF, our VDM-rendering produces physically plausible dynamics of hair and earrings, as well as complex light interactions like shadows, reflections and refractions. Crucially, we show that common 3D representaion (e.g. 3DGS, NeRF) are not necessary for photorealistic and view-point controllable avatars. VDA is quite affordable for the research community, since training only requires 34 A100-GPU hours on public multi-view datasets, yet VDA generalizes far beyond them: to paintings, human-like creatures and full 360° orbits. 🌐 ▶️ Great work by Simon Giebenhain, Tobias Kirschstein, Liam Schoneveld, Davide Davoli, Zhe Chen

Matthias Niessner

17,715 次观看 • 4 天前

📢Introducing 360Anything, our method for lifting any perspective image or video to gravity-aligned 360° panoramas without using any camera or 3D information. This enables consistent novel view synthesis and 3D scene reconstruction. Project page: 🧵

📢Introducing 360Anything, our method for lifting any perspective image or video to gravity-aligned 360° panoramas without using any camera or 3D information. This enables consistent novel view synthesis and 3D scene reconstruction. Project page: 🧵

Ziyi Wu

63,573 次观看 • 8 个月前

Introducing Higgsfield 3D Jutsu, powered by GPT-6 Astra. Turn a single prompt into a fully editable, interactive 3D world. > Generate complex 3D scenes from text > Control geometry, layout, and details > Refine environment instantly with simple prompts Now live on Higgsfield.

Introducing Higgsfield 3D Jutsu, powered by GPT-6 Astra. Turn a single prompt into a fully editable, interactive 3D world. > Generate complex 3D scenes from text > Control geometry, layout, and details > Refine environment instantly with simple prompts Now live on Higgsfield.

Higgsfield AI 🧩

99,987 次观看 • 1 个月前

A test of real-time transfer of images captured on Quest to a PC, where they are converted into 3D using Apple's SHARP. Currently, the final 3DGS output is displayed on the PC. The next step is to bring the result back into Quest for in-headset viewing. #3DGS #metaquest #AR

A test of real-time transfer of images captured on Quest to a PC, where they are converted into 3D using Apple's SHARP. Currently, the final 3DGS output is displayed on the PC. The next step is to bring the result back into Quest for in-headset viewing. #3DGS #metaquest #AR

Takashi Yoshinaga

11,575 次观看 • 6 个月前

VAST AI releases Triplane Meets Gaussian Splatting on Hugging Face Fast and Generalizable Single-View 3D Reconstruction with Transformers demo: TGS enables fast reconstruction from single-view image. It builds the 3D representation upon a hybrid Triplane-Gaussian representation by evaluating a transformer-based framework, from which 3D Gaussians would be decoded

VAST AI releases Triplane Meets Gaussian Splatting on Hugging Face Fast and Generalizable Single-View 3D Reconstruction with Transformers demo: TGS enables fast reconstruction from single-view image. It builds the 3D representation upon a hybrid Triplane-Gaussian representation by evaluating a transformer-based framework, from which 3D Gaussians would be decoded

AK

113,242 次观看 • 2 年前

Learning Temporally Consistent Video Depth from Video Diffusion Priors This work addresses the challenge of video depth estimation, which expects not only per-frame accuracy but, more importantly, cross-frame consistency. Instead of directly

Learning Temporally Consistent Video Depth from Video Diffusion Priors This work addresses the challenge of video depth estimation, which expects not only per-frame accuracy but, more importantly, cross-frame consistency. Instead of directly

AK

81,666 次观看 • 2 年前

𝗜𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝗶𝗻𝗴 𝗨𝗩𝗚𝗦 We introduce 𝗨𝗩𝗚𝗦, a new 2D representation of 3D Gaussian Splatting (3DGS) that leverages spherical mapping. Website: Paper:

𝗜𝗻𝘁𝗿𝗼𝗱𝘂𝗰𝗶𝗻𝗴 𝗨𝗩𝗚𝗦 We introduce 𝗨𝗩𝗚𝗦, a new 2D representation of 3D Gaussian Splatting (3DGS) that leverages spherical mapping. Website: Paper:

Nikolaos Sarafianos

10,753 次观看 • 1 年前

(1/2) 📢📢𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧𝐀𝐯𝐚𝐭𝐚𝐫𝐬 📢📢 High-fidelity 3D head avatars with precise control over viewpoint, expression, and pose. -> Our parametric 3D model enables control & consistency + 2D diffusion makes it photoreal.

(1/2) 📢📢𝐃𝐢𝐟𝐟𝐮𝐬𝐢𝐨𝐧𝐀𝐯𝐚𝐭𝐚𝐫𝐬 📢📢 High-fidelity 3D head avatars with precise control over viewpoint, expression, and pose. -> Our parametric 3D model enables control & consistency + 2D diffusion makes it photoreal.

Matthias Niessner

66,436 次观看 • 2 年前