Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Introducing CUT3R! An online 3D reasoning framework for many 3D tasks directly from just RGB. For static or dynamic scenes. Video or image collections, all in one!

Qianqian Wang

4,566 subscribers

145,864 views • 1 year ago •via X (Twitter)

Science & Technology

Anya Rossi• Live Now

Private livecam show

10 Comments

Qianqian Wang1 year ago

Key insight is to store 3D information in a latent persistent state that gets continuously updated as images come in. This enables it to continuously refine its predictions as it sees more images!

Qianqian Wang1 year ago

From the state you can also predict unseen parts. This also improves as you see more. We have a lot of results on the project site: check it out!

Qianqian Wang1 year ago

Thanks to the amazing team @Ify_Zhang @holynski_, Alexei Efros, and @akanazawa! Code and model will be released soon!

RedDeer.Games4 years ago

Customise the colours yourself or simply connect the Joy-Cons ! Head to ➡️ for more details about #AAAClock ! #indiedev #indiegame #IndieGameDev #nintendo #OLED #switchOLED

Vikas Reddy1 year ago

If you have multiple cameras w/ accurate timestamps does this work still? Seems like it would? Excited to try it out either way!

Tom Boyle1 year ago

CUT3R looks like a breakthrough for 3D reasoning! Being able to handle both static and dynamic scenes is incredible. What’s your favorite use case so far?

Swirls Defy1 year ago

Tagging @accursedfarms wrt magic software

Miguel Hernández1 year ago

CUT3R looks like a game-changer for 3D reasoning tasks! Can you share some examples of how it's been used in real-world applications?

buffaly1 year ago

@berkeley_ai Fantastic work! I'd love to get pointed to the paper to learn more

Yixuan Wang1 year ago

Very amazing work! If the input camera is dynamic and contains dynamic objects in the scene, will the reconstructed scene retain the information from past frames, like Dynamic SLAM? I checked out the dynamic demo on the website, but they seem to focus on current frame mainly?

Related Videos

🌟Your static 3D world models are now alive and interactable! 🚀Introducing NeuROK, a neural simulation framework that turns any static 3D object into an interactive 4D asset — no per-category physics, no physical annotations for training. 📄 🧵 1/n

🌟Your static 3D world models are now alive and interactable! 🚀Introducing NeuROK, a neural simulation framework that turns any static 3D object into an interactive 4D asset — no per-category physics, no physical annotations for training. 📄 🧵 1/n

Chen Geng

31,336 views • 14 days ago

📢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

46,859 views • 4 months ago

Convert any image into an animated 3D model for your web or app design!

Convert any image into an animated 3D model for your web or app design!

Jan Mraz

41,270 views • 7 months ago

Motion cues from human videos + Reasoning from VLMs enables 3D hand trajectory prediction in-the-wild for novel tasks in novel scenes! Very excited to share the EgoMan project, a huge effort led by Mingfei Chen during an internship with us at Meta. Video summary ⬇️

Motion cues from human videos + Reasoning from VLMs enables 3D hand trajectory prediction in-the-wild for novel tasks in novel scenes! Very excited to share the EgoMan project, a huge effort led by Mingfei Chen during an internship with us at Meta. Video summary ⬇️

Homanga Bharadhwaj @ CVPR

10,506 views • 5 months ago

3D-R1 is out on Hugging Face Enhancing Reasoning in 3D VLMs for Unified Scene Understanding 3D-R1 is an open-source generalist model that enhances the reasoning of 3D VLMs for unified scene understanding

3D-R1 is out on Hugging Face Enhancing Reasoning in 3D VLMs for Unified Scene Understanding 3D-R1 is an open-source generalist model that enhances the reasoning of 3D VLMs for unified scene understanding

AK

34,729 views • 10 months ago

Spatial reasoning is a major challenge for the foundation models today, even in simple tasks like arranging objects in 3D space. #CVPR2025 Introducing LayoutVLM, a differentiable optimization framework that uses VLM to spatially reason about diverse scene layouts from unlabeled assets and open-ended language instructions 1/n

Spatial reasoning is a major challenge for the foundation models today, even in simple tasks like arranging objects in 3D space. #CVPR2025 Introducing LayoutVLM, a differentiable optimization framework that uses VLM to spatially reason about diverse scene layouts from unlabeled assets and open-ended language instructions 1/n

Fan-Yun Sun

92,541 views • 1 year ago

#ICCV2025 🤩3D world generation is cool, but it is cooler to play with the worlds using 3D actions 👆💨, and see what happens! — Introducing *WonderPlay*: Now you can create dynamic 3D scenes that respond to your 3D actions from a single image! Web: 🧵1/7

#ICCV2025 🤩3D world generation is cool, but it is cooler to play with the worlds using 3D actions 👆💨, and see what happens! — Introducing WonderPlay: Now you can create dynamic 3D scenes that respond to your 3D actions from a single image! Web: 🧵1/7

Hong-Xing (Koven) Yu

57,590 views • 11 months ago

⚡️ Introducing Bolt3D ⚡️ Bolt3D generates interactive 3D scenes in less than 7 seconds on a single GPU from one or more images. It features a latent diffusion model that *directly* generates 3D Gaussians of seen and unseen regions, without any test time optimization. 🧵👇 (1/9)

⚡️ Introducing Bolt3D ⚡️ Bolt3D generates interactive 3D scenes in less than 7 seconds on a single GPU from one or more images. It features a latent diffusion model that directly generates 3D Gaussians of seen and unseen regions, without any test time optimization. 🧵👇 (1/9)

Stan Szymanowicz

125,828 views • 1 year ago

What if we can simulate an *interactive 3D world*, from a single image, in the wild, in real time? Introducing PointWorld-1B: a large pre-trained 3D world model that predicts env dynamics given RGB-D capture and robot actions. 🌐 from Stanford University NVIDIA

What if we can simulate an interactive 3D world, from a single image, in the wild, in real time? Introducing PointWorld-1B: a large pre-trained 3D world model that predicts env dynamics given RGB-D capture and robot actions. 🌐 from Stanford University NVIDIA

Wenlong Huang

272,592 views • 5 months ago

🌟 Create anything in 3D! 🌟 Introducing CAT3D: a new method that generates high-fidelity 3D scenes from any number of real or generated images in one minute, powered by multi-view diffusion models. w/ lovely coauthors Aleksander Holynski, Ben Poole and an amazing team!

🌟 Create anything in 3D! 🌟 Introducing CAT3D: a new method that generates high-fidelity 3D scenes from any number of real or generated images in one minute, powered by multi-view diffusion models. w/ lovely coauthors Aleksander Holynski, Ben Poole and an amazing team!

Ruiqi Gao

152,867 views • 2 years ago

Phidias A Generative Model for Creating 3D Content from Text, Image, and 3D Conditions with Reference-Augmented Diffusion discuss: In 3D modeling, designers often use an existing 3D model as a reference to create new ones. This practice has inspired the development of Phidias, a novel generative model that uses diffusion for reference-augmented 3D generation. Given an image, our method leverages a retrieved or user-provided 3D reference model to guide the generation process, thereby enhancing the generation quality, generalization ability, and controllability. Our model integrates three key components: 1) meta-ControlNet that dynamically modulates the conditioning strength, 2) dynamic reference routing that mitigates misalignment between the input image and 3D reference, and 3) self-reference augmentations that enable self-supervised training with a progressive curriculum. Collectively, these designs result in a clear improvement over existing methods. Phidias establishes a unified framework for 3D generation using text, image, and 3D conditions with versatile applications.

Phidias A Generative Model for Creating 3D Content from Text, Image, and 3D Conditions with Reference-Augmented Diffusion discuss: In 3D modeling, designers often use an existing 3D model as a reference to create new ones. This practice has inspired the development of Phidias, a novel generative model that uses diffusion for reference-augmented 3D generation. Given an image, our method leverages a retrieved or user-provided 3D reference model to guide the generation process, thereby enhancing the generation quality, generalization ability, and controllability. Our model integrates three key components: 1) meta-ControlNet that dynamically modulates the conditioning strength, 2) dynamic reference routing that mitigates misalignment between the input image and 3D reference, and 3) self-reference augmentations that enable self-supervised training with a progressive curriculum. Collectively, these designs result in a clear improvement over existing methods. Phidias establishes a unified framework for 3D generation using text, image, and 3D conditions with versatile applications.

AK

25,120 views • 1 year ago

ChatGPT Image arrived to Krea. edit images in Chat, render 3D scenes in Stage, or generate images from hundreds of Krea styles. more coming soon!

ChatGPT Image arrived to Krea. edit images in Chat, render 3D scenes in Stage, or generate images from hundreds of Krea styles. more coming soon!

KREA AI

112,657 views • 1 year ago

Quickly create real-time dance content that you can choose to use directly for virtual live streaming or interactive experiences in VR. All it takes is capturing with a webcam, powered by an AI-generated 3D motion solution. 3D models from Warudo @ Hakuya Labs

Quickly create real-time dance content that you can choose to use directly for virtual live streaming or interactive experiences in VR. All it takes is capturing with a webcam, powered by an AI-generated 3D motion solution. 3D models from Warudo @ Hakuya Labs

CYANPUPPETS

16,462 views • 2 years ago

Generate 3D from just one image, on your computer. franzipol has built a gradio web ui for Apple's Sharp, which infers an entire 3D view from just one image. You can run it on any computer with 1-click: Mac, Linux, Windows. Just need 10GB VRAM (16GB memory on Macs).

Generate 3D from just one image, on your computer. franzipol has built a gradio web ui for Apple's Sharp, which infers an entire 3D view from just one image. You can run it on any computer with 1-click: Mac, Linux, Windows. Just need 10GB VRAM (16GB memory on Macs).

cocktail peanut

157,279 views • 5 months ago

🚨 Paper Alert Our recent breakthrough CAST: Component-Aligned 3D Scene Reconstruction from an RGB Image has been accepted by ACM SIGGRAPH 2025 Journal Track! CAST will change the way create scenes in 3D Art and Embody AI. 🚀Soon available at 👇Details

🚨 Paper Alert Our recent breakthrough CAST: Component-Aligned 3D Scene Reconstruction from an RGB Image has been accepted by ACM SIGGRAPH 2025 Journal Track! CAST will change the way create scenes in 3D Art and Embody AI. 🚀Soon available at 👇Details

Deemos

34,110 views • 1 year ago

⚡️Your video → a 3D scene. Just in seconds.⚡️ Only *one* model. No more steps❌. Just results🔝. #WorldMirror reconstructs everything (3DGS, depth, cameras) from any inputs (image, video, 3D prior) *all-at-once*! code: arxiv:

⚡️Your video → a 3D scene. Just in seconds.⚡️ Only one model. No more steps❌. Just results🔝. #WorldMirror reconstructs everything (3DGS, depth, cameras) from any inputs (image, video, 3D prior) all-at-once! code: arxiv:

Tengfei Wang

66,897 views • 7 months ago

🥳 Nano Banana is available for image generation in Meshy! Text → Image → 3D. Try the all-in-one AI workflow now!

🥳 Nano Banana is available for image generation in Meshy! Text → Image → 3D. Try the all-in-one AI workflow now!

MeshyAI

34,732 views • 9 months ago

You can build your 3D models directly from a prompt into Claude code. All thanks to this beautiful python project, build123 a parametric modeling framework for 2D and 3D CAD. It is built on the Open Cascade geometric kernel and it is suitable for 3D printing.

You can build your 3D models directly from a prompt into Claude code. All thanks to this beautiful python project, build123 a parametric modeling framework for 2D and 3D CAD. It is built on the Open Cascade geometric kernel and it is suitable for 3D printing.

Marco Franzon

20,743 views • 2 months ago

SceneDreamer: Unbounded 3D Scene Generation from 2D Image Collections abs: project page:

SceneDreamer: Unbounded 3D Scene Generation from 2D Image Collections abs: project page:

AK

317,619 views • 3 years ago