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.

Given a monocular video as input, #HOLD reconstructs 3D hand and object surfaces for every frame without assuming a known object template. Our key insight is that interacting hands and objects provide complementary cues about each other's shape and pose. 1/4

Michael Black

97,521 subscribers

21,594 views • 2 years ago •via X (Twitter)

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Monocular pose estimation has gotten really good Grab any 2D video and transfer the performance to a 3D character

Monocular pose estimation has gotten really good Grab any 2D video and transfer the performance to a 3D character

Bilawal Sidhu

27,074 views • 1 year ago

As a Product Designer with an educational and professional background in stop motion and frame-to-frame animation, creating these loaders was seamless. I carefully considered the shape of each frame for the desired animation. Here's my process: 👇👇 ◉ ○ ○ ○ - 1/4

As a Product Designer with an educational and professional background in stop motion and frame-to-frame animation, creating these loaders was seamless. I carefully considered the shape of each frame for the desired animation. Here's my process: 👇👇 ◉ ○ ○ ○ - 1/4

Muhammed Adepoju - Design Yoda

41,672 views • 2 years ago

Grasps are one of the primary ways in which we interact with and shape our environments. How can we faithfully capture human grasps with details such as hand/object shape and contact points? At #CVPR2026, we present MANUS, a method to accurately reconstruct grasps and contacts. 🧵

Grasps are one of the primary ways in which we interact with and shape our environments. How can we faithfully capture human grasps with details such as hand/object shape and contact points? At #CVPR2026, we present MANUS, a method to accurately reconstruct grasps and contacts. 🧵

Srinath Sridhar

10,767 views • 2 years ago

GSTAR: Gaussian Surface Tracking and Reconstruction Contributions: • A new framework for tracking and reconstructing dynamic scenes, combining 3D Gaussians and meshes to effectively manage changes in topology. • A method for Gaussian unbinding and surface re-meshing, allowing for the generation of new surfaces as topologies evolve. • A method for handling large or fast deformations of surfaces between frames using scene flow warping. Abstract (excerpt): However, tracking dynamic surfaces with 3D Gaussians remains challenging due to complex topology changes, such as surfaces appearing, disappearing, or splitting. To address these challenges, we propose GSTAR, a novel method that achieves photo-realistic rendering, accurate surface reconstruction, and reliable 3D tracking for general dynamic scenes with changing topology. Given multi-view captures as input, GSTAR binds Gaussians to mesh faces to represent dynamic objects. For surfaces with consistent topology, GSTAR maintains the mesh topology and tracks the meshes using Gaussians.

GSTAR: Gaussian Surface Tracking and Reconstruction Contributions: • A new framework for tracking and reconstructing dynamic scenes, combining 3D Gaussians and meshes to effectively manage changes in topology. • A method for Gaussian unbinding and surface re-meshing, allowing for the generation of new surfaces as topologies evolve. • A method for handling large or fast deformations of surfaces between frames using scene flow warping. Abstract (excerpt): However, tracking dynamic surfaces with 3D Gaussians remains challenging due to complex topology changes, such as surfaces appearing, disappearing, or splitting. To address these challenges, we propose GSTAR, a novel method that achieves photo-realistic rendering, accurate surface reconstruction, and reliable 3D tracking for general dynamic scenes with changing topology. Given multi-view captures as input, GSTAR binds Gaussians to mesh faces to represent dynamic objects. For surfaces with consistent topology, GSTAR maintains the mesh topology and tracks the meshes using Gaussians.

MrNeRF

22,698 views • 1 year ago

Another explanation for Oumuamua's unusual shape has been proposed The asteroid Oumuamua, which flew through the Solar System in 2017, amazed scientists with its unprecedentedly elongated shape. It is 230 meters long and about 35 meters wide. Astronomers have never seen such objects before and tried to explain its origin. Oumuamua is the first known object to arrive from interstellar space. Its trajectory and speed left no doubt that it does not belong to the Solar System. The object rotated around its axis, changing its brightness, which made it possible to determine its size.

Another explanation for Oumuamua's unusual shape has been proposed The asteroid Oumuamua, which flew through the Solar System in 2017, amazed scientists with its unprecedentedly elongated shape. It is 230 meters long and about 35 meters wide. Astronomers have never seen such objects before and tried to explain its origin. Oumuamua is the first known object to arrive from interstellar space. Its trajectory and speed left no doubt that it does not belong to the Solar System. The object rotated around its axis, changing its brightness, which made it possible to determine its size.

Black Hole

77,373 views • 1 year ago

Another explanation for Oumuamua's unusual shape has been proposed The asteroid Oumuamua, which flew through the Solar System in 2017, amazed scientists with its unprecedentedly elongated shape. It is 230 meters long and about 35 meters wide. Astronomers have never seen such objects before and tried to explain its origin. Oumuamua is the first known object to arrive from interstellar space. Its trajectory and speed left no doubt that it does not belong to the Solar System. The object rotated around its axis, changing its brightness, which made it possible to determine its size.

Another explanation for Oumuamua's unusual shape has been proposed The asteroid Oumuamua, which flew through the Solar System in 2017, amazed scientists with its unprecedentedly elongated shape. It is 230 meters long and about 35 meters wide. Astronomers have never seen such objects before and tried to explain its origin. Oumuamua is the first known object to arrive from interstellar space. Its trajectory and speed left no doubt that it does not belong to the Solar System. The object rotated around its axis, changing its brightness, which made it possible to determine its size.

Black Hole

237,039 views • 1 year ago

Manage your 3D objects efficiently with Object List! With this handy tool, you can select, delete, and duplicate and group several 3D items together and adjust the light source settings, all while working with multiple 3D objects on the same layer. #clipstudio

Manage your 3D objects efficiently with Object List! With this handy tool, you can select, delete, and duplicate and group several 3D items together and adjust the light source settings, all while working with multiple 3D objects on the same layer. #clipstudio

CLIP STUDIO PAINT

11,764 views • 6 months ago

We present “3D magician”: TADA! Text to Animatable Digital Avatars. Given a textual description as input only, our method TADA generates expressive animatable 3D avatars with high-quality geometry and lifelike textures. (1/10)

We present “3D magician”: TADA! Text to Animatable Digital Avatars. Given a textual description as input only, our method TADA generates expressive animatable 3D avatars with high-quality geometry and lifelike textures. (1/10)

Hongwei Yi

52,306 views • 2 years ago

NVIDIA finally released Neuralangelo's source code! The model can turn videos from any device into detailed 3D structures, fully replicating buildings, sculptures, or other real aworld objects or spaces virtually. Here's how it works: A model utilizes a 2D video with multiple angles of an object or scene. I selects frames from different viewpoints to understand depth, size, and shape. The AI creates an initial 3D representation, similar to a sculptor shaping a subject. The render is optimized to enhance details, like a sculptor refining texture. The outcome is a 3D object or scene suitable for virtual reality, digital twins, or robotics.

NVIDIA finally released Neuralangelo's source code! The model can turn videos from any device into detailed 3D structures, fully replicating buildings, sculptures, or other real aworld objects or spaces virtually. Here's how it works: A model utilizes a 2D video with multiple angles of an object or scene. I selects frames from different viewpoints to understand depth, size, and shape. The AI creates an initial 3D representation, similar to a sculptor shaping a subject. The render is optimized to enhance details, like a sculptor refining texture. The outcome is a 3D object or scene suitable for virtual reality, digital twins, or robotics.

Lior Alexander

478,018 views • 2 years ago

Boys invented arm wrestling so they could hold hands and look in each other's eyes

Boys invented arm wrestling so they could hold hands and look in each other's eyes

WeirdHumanBeing

13,905 views • 10 months ago

Gemma 4 just dropped. I had it captioning video in real-time within an hour. Running locally on a MacBook. No cloud. No API. Real-time scene understanding. Oh and SAM3 is segmenting every object in the same frame. Same laptop.

Gemma 4 just dropped. I had it captioning video in real-time within an hour. Running locally on a MacBook. No cloud. No API. Real-time scene understanding. Oh and SAM3 is segmenting every object in the same frame. Same laptop.

Maziyar PANAHI

196,846 views • 2 months ago

improving my 3D hand tracking template by adding depth-anything-v2 using mediapipe for X/Y movement and depth-anything for Z movement here i'm controlling a threejs scene in realtime with input from a single webcam

improving my 3D hand tracking template by adding depth-anything-v2 using mediapipe for X/Y movement and depth-anything for Z movement here i'm controlling a threejs scene in realtime with input from a single webcam

AA

28,216 views • 2 months ago

If you’re not in freefall, you’re likely contacting something. Yet 3D human-object interaction (HOI) reconstruction remains underexplored. PICO (#CVPR2025) recovers humans 🏃‍♂️, objects 🏓, and their interactions 👉🍎- all in 3D, from just a single internet image. 1/11

If you’re not in freefall, you’re likely contacting something. Yet 3D human-object interaction (HOI) reconstruction remains underexplored. PICO (#CVPR2025) recovers humans 🏃‍♂️, objects 🏓, and their interactions 👉🍎- all in 3D, from just a single internet image. 1/11

Shashank Tripathi

24,114 views • 1 year ago

This 3d hand is 100% made in Spline The possibilities of the new "Shape Blend" tool are incredible! Objects become more and more complex and realistic. Remix it:

This 3d hand is 100% made in Spline The possibilities of the new "Shape Blend" tool are incredible! Objects become more and more complex and realistic. Remix it:

Max

15,233 views • 1 year ago

Using /goal and ForgeCAD, Codex is now able to reconstruct a CAD object from video:

Using /goal and ForgeCAD, Codex is now able to reconstruct a CAD object from video:

Ruben Kostandyan

21,872 views • 1 month ago

Massive performance improvement. This is a bit more of technical post, but man do I love this stuff! Units navigate the map using a 'Navigation Mesh'. Before, I was using one giant nav mesh that spanned the entire map. The more objects that were placed (especially on a large map such as this 'RadarAttack' map designed by Syphotic | Steel Command), the larger the 'lag' would be after placement. You can see here that there is a massive frame drop and the navmesh doesnt update for almost 5 seconds. Now, there are a ton of tiny navmeshes that connect to one another, and together they cover the whole map. Now, when an object is placed, the navmesh will update instantly, because it no longer needs to parse through every object on the map (potentially thousands!!!). It only needs to parse through the objects that exist in the mini navmesh that the object was placed in (probably only 1-5 objects now!). Performance XP Boost +100! Charles Horwood You might appreciate this one :) #Rts #RTSGame #IndieGame

Massive performance improvement. This is a bit more of technical post, but man do I love this stuff! Units navigate the map using a 'Navigation Mesh'. Before, I was using one giant nav mesh that spanned the entire map. The more objects that were placed (especially on a large map such as this 'RadarAttack' map designed by Syphotic | Steel Command), the larger the 'lag' would be after placement. You can see here that there is a massive frame drop and the navmesh doesnt update for almost 5 seconds. Now, there are a ton of tiny navmeshes that connect to one another, and together they cover the whole map. Now, when an object is placed, the navmesh will update instantly, because it no longer needs to parse through every object on the map (potentially thousands!!!). It only needs to parse through the objects that exist in the mini navmesh that the object was placed in (probably only 1-5 objects now!). Performance XP Boost +100! Charles Horwood You might appreciate this one :) #Rts #RTSGame #IndieGame

Smitty | Steel Command

60,072 views • 5 months ago

Introducing 📦𝗔𝗿𝘁𝗶𝗟𝗮𝘁𝗲𝗻𝘁🔧 (SIGGRAPH Asia 2025) — a high-quality 3D diffusion model that explicitly models object articulation, paving the way for richer, more realistic assets in embodied AI and simulation: – Generates fully articulated 3D objects – Physically plausible joints & motion – High-fidelity 3D Gaussian appearance – Supports generation from a single real image arXiv: Project: Code (coming soon):

Introducing 📦𝗔𝗿𝘁𝗶𝗟𝗮𝘁𝗲𝗻𝘁🔧 (SIGGRAPH Asia 2025) — a high-quality 3D diffusion model that explicitly models object articulation, paving the way for richer, more realistic assets in embodied AI and simulation: – Generates fully articulated 3D objects – Physically plausible joints & motion – High-fidelity 3D Gaussian appearance – Supports generation from a single real image arXiv: Project: Code (coming soon):

Xingang Pan

11,473 views • 7 months ago

In Figma Design you can easily cut up a shape with a path using the shape builder tool, without moving over to draw! 🤯 Just draw the path you want to use to "Cut." Just select your object and path and press "M," and click your regions. #FigmaTip

In Figma Design you can easily cut up a shape with a path using the shape builder tool, without moving over to draw! 🤯 Just draw the path you want to use to "Cut." Just select your object and path and press "M," and click your regions. #FigmaTip

miggi from figma

243,258 views • 9 months ago

WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments TL;DR: We present WildGS-SLAM, a robust monocular RGB SLAM system. - Utilizes uncertainty-aware tracking and mapping to handle dynamic scenes. - Leverages DINOv2-based uncertainty maps for dynamic object removal. - Improves tracking and mapping. - Enables high-quality view synthesis.

WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments TL;DR: We present WildGS-SLAM, a robust monocular RGB SLAM system. - Utilizes uncertainty-aware tracking and mapping to handle dynamic scenes. - Leverages DINOv2-based uncertainty maps for dynamic object removal. - Improves tracking and mapping. - Enables high-quality view synthesis.

MrNeRF

17,382 views • 1 year ago

I always thought camera pose estimation is necessary for 3D reconstruction, until Zequn and Stephen proved me wrong! Introducing PreF3R, purely feed-forward 3D Gaussian Splatting without any intermediate pose estimation and COLMAP initialization. Video in, 3D Gaussians and novel view rendering out. Key technique: spatial memory network from Spann3R and Gaussian head supervised by pointmap loss plus photometric loss. See more at:

I always thought camera pose estimation is necessary for 3D reconstruction, until Zequn and Stephen proved me wrong! Introducing PreF3R, purely feed-forward 3D Gaussian Splatting without any intermediate pose estimation and COLMAP initialization. Video in, 3D Gaussians and novel view rendering out. Key technique: spatial memory network from Spann3R and Gaussian head supervised by pointmap loss plus photometric loss. See more at:

Heng Yang

20,160 views • 1 year ago