Loading video...

Video Failed to Load

Go Home

We present EgoReAct: Real-time 3D human reaction generation from streaming egocentric video. 🌟Reacting to streaming egocentric video is something humans do every day. We hope EgoReAct makes human motion more human-like. 🔎 What we found: existing ego-reaction data can be spatially inconsistent (e.g., moving reactions paired with fixed-camera videos),...

frankzydou's profile picture

Zhiyang (Frank) Dou

2,360 subscribers

11,427 views • 9 months ago •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

8 Comments

Zhiyang (Frank) Dou's profile picture
Zhiyang (Frank) Dou9 months ago

typo: paris->pairs.

Kavin Raj's profile picture
Kavin Raj9 months ago

crazy implementation - love it!

Andres Rey's profile picture
Andres Rey9 months ago

EgoReAct solving ego-reaction generation from streaming video is brilliant. Using existing ego-reaction data = building on proven patterns. This will transform VR avatars, motion capture, and digital humans. Human-like motion is now generatable at scale, not handcrafted.

crisiumnih's profile picture
crisiumnih9 months ago

This is such a abstract idea lmao. Makes sense using a planner and then generation. How long it generates till the architecture blows up? Great work though!

Mentis 🇦🇺's profile picture
Mentis 🇦🇺9 months ago

Hardware companions always feel like a solution looking for a problem.

Zhiyang (Frank) Dou's profile picture
Zhiyang (Frank) Dou9 months ago

Thanks for your comment! Just to clarify, EgoReAct is entirely software-based and focuses on real-time 3D human reaction generation from streaming egocentric video (the video can be from a common video recording device). We’re working on improving the generation of human reaction from egocentric video streams, which can be a good tool for AR/VR, animation, robotics, and beyond.😊

Mentis 🇦🇺's profile picture
Mentis 🇦🇺9 months ago

Thanks yeah - it's brilliant. The point is the hardware/physical layer is such a limiter... you're going the right way!

Zhiyang (Frank) Dou's profile picture
Zhiyang (Frank) Dou9 months ago

🙇‍Appreciate it! Yes—I think standard RGB cameras already work well for our setting, but better devices will always be helpful! Besides, we definitely need to further scale the model up for stronger generalization.

Related Videos

Excited to share our latest work on 🎧spatial audio-driven human motion generation. We aim to tackle a largely underexplored yet important problem of enabling virtual humans to move naturally in response to spatial audio—capturing not just what is heard, but also where the sound is coming from. To this end, we introduce the Spatial Audio-Driven Human Motion (SAM) dataset—the first comprehensive dataset featuring paired high-quality human motion and spatial audio recordings. For benchmarking, we develop a generative framework for human MOtion generation driven by SPAtial audio, termed MOSPA, which learns to synthesize realistic and diverse human motions conditioned on spatial audio input. We hope this research could provide a foundation for future research in spatial perception, virtual characters, and embodied AI. The dataset and model will be open-sourced soon. A big thank you to our intern, Shuyang Xu, for the wonderful collaboration! Congratulations, Shuyang! Project page: Paper: Video: #Animation #CG #CV #AIGC #DL #Deeplearning #Motion #Graphics #AI #GenerativeAI
3:58

Excited to share our latest work on 🎧spatial audio-driven human motion generation. We aim to tackle a largely underexplored yet important problem of enabling virtual humans to move naturally in response to spatial audio—capturing not just what is heard, but also where the sound is coming from. To this end, we introduce the Spatial Audio-Driven Human Motion (SAM) dataset—the first comprehensive dataset featuring paired high-quality human motion and spatial audio recordings. For benchmarking, we develop a generative framework for human MOtion generation driven by SPAtial audio, termed MOSPA, which learns to synthesize realistic and diverse human motions conditioned on spatial audio input. We hope this research could provide a foundation for future research in spatial perception, virtual characters, and embodied AI. The dataset and model will be open-sourced soon. A big thank you to our intern, Shuyang Xu, for the wonderful collaboration! Congratulations, Shuyang! Project page: Paper: Video: #Animation #CG #CV #AIGC #DL #Deeplearning #Motion #Graphics #AI #GenerativeAI

Zhiyang (Frank) Dou

14,728 views • 1 year ago

🔥 Introducing MVLift: Generate realistic 3D motion without any 3D training data - just using 2D poses from monocular videos! Applicable to human motion, human-object interaction & animal motion. Joint work w/ Jiajun Wu & Karen 💡 How? We reformulate 3D motion estimation as generating consistent multi-view 2D pose sequences. Our framework uses 2D motion diffusion to progressively establish multi-view consistency, requiring only single-view 2D pose sequences for training. Project: Video with demonstration: Paper:
0:21

🔥 Introducing MVLift: Generate realistic 3D motion without any 3D training data - just using 2D poses from monocular videos! Applicable to human motion, human-object interaction & animal motion. Joint work w/ Jiajun Wu & Karen 💡 How? We reformulate 3D motion estimation as generating consistent multi-view 2D pose sequences. Our framework uses 2D motion diffusion to progressively establish multi-view consistency, requiring only single-view 2D pose sequences for training. Project: Video with demonstration: Paper:

Jiaman Li

15,886 views • 1 year ago

EgoLifter Open-world 3D Segmentation for Egocentric Perception In this paper we present EgoLifter, a novel system that can automatically segment scenes captured from egocentric sensors into a complete decomposition of individual 3D objects. The system is specifically
0:20

EgoLifter Open-world 3D Segmentation for Egocentric Perception In this paper we present EgoLifter, a novel system that can automatically segment scenes captured from egocentric sensors into a complete decomposition of individual 3D objects. The system is specifically

AK

41,117 views • 2 years ago

Imitation learning has a data scarcity problem. Introducing EgoDex from Apple, the largest and most diverse dataset of dexterous human manipulation to date — 829 hours of egocentric video + paired 3D hand poses across 194 tasks. Now on arxiv: (1/4)
0:27

Imitation learning has a data scarcity problem. Introducing EgoDex from Apple, the largest and most diverse dataset of dexterous human manipulation to date — 829 hours of egocentric video + paired 3D hand poses across 194 tasks. Now on arxiv: (1/4)

Ryan Hoque

114,335 views • 1 year ago

PointOdyssey: A Large-Scale Synthetic Dataset for Long-Term Point Tracking paper page: introduce PointOdyssey, a large-scale synthetic dataset, and data generation framework, for the training and evaluation of long-term fine-grained tracking algorithms. Our goal is to advance the state-of-the-art by placing emphasis on long videos with naturalistic motion. Toward the goal of naturalism, we animate deformable characters using real-world motion capture data, we build 3D scenes to match the motion capture environments, and we render camera viewpoints using trajectories mined via structure-from-motion on real videos. We create combinatorial diversity by randomizing character appearance, motion profiles, materials, lighting, 3D assets, and atmospheric effects. Our dataset currently includes 104 videos, averaging 2,000 frames long, with orders of magnitude more correspondence annotations than prior work. We show that existing methods can be trained from scratch in our dataset and outperform the published variants. Finally, we introduce modifications to the PIPs point tracking method, greatly widening its temporal receptive field, which improves its performance on PointOdyssey as well as on two real-world benchmarks.
0:33

PointOdyssey: A Large-Scale Synthetic Dataset for Long-Term Point Tracking paper page: introduce PointOdyssey, a large-scale synthetic dataset, and data generation framework, for the training and evaluation of long-term fine-grained tracking algorithms. Our goal is to advance the state-of-the-art by placing emphasis on long videos with naturalistic motion. Toward the goal of naturalism, we animate deformable characters using real-world motion capture data, we build 3D scenes to match the motion capture environments, and we render camera viewpoints using trajectories mined via structure-from-motion on real videos. We create combinatorial diversity by randomizing character appearance, motion profiles, materials, lighting, 3D assets, and atmospheric effects. Our dataset currently includes 104 videos, averaging 2,000 frames long, with orders of magnitude more correspondence annotations than prior work. We show that existing methods can be trained from scratch in our dataset and outperform the published variants. Finally, we introduce modifications to the PIPs point tracking method, greatly widening its temporal receptive field, which improves its performance on PointOdyssey as well as on two real-world benchmarks.

AK

122,567 views • 3 years ago

Most egocentric data at scale is private. We think that's worth changing. With Hugging Face, we've released the first 10,000 hours of EgoSuite-Open100K — an open egocentric human dataset for Physical AI. #Robotics
0:50

Most egocentric data at scale is private. We think that's worth changing. With Hugging Face, we've released the first 10,000 hours of EgoSuite-Open100K — an open egocentric human dataset for Physical AI. #Robotics

Lightwheel

11,755 views • 2 months ago

We present VLM-3R: a Vision-Language Model capable of 3D spatial reasoning from monocular video, grounding visual cues, geometry, and camera motion. ✅ No depth sensor ✅ No pre-built 3D maps ✅ End-to-end spatial + temporal reasoning 🔗 Code & benchmark: #VLM #3DVision #LLMs
0:20

We present VLM-3R: a Vision-Language Model capable of 3D spatial reasoning from monocular video, grounding visual cues, geometry, and camera motion. ✅ No depth sensor ✅ No pre-built 3D maps ✅ End-to-end spatial + temporal reasoning 🔗 Code & benchmark: #VLM #3DVision #LLMs

Zhiwen(Aaron) Fan

14,895 views • 1 year ago

ReViV: Reconstructing the Viewer and the View in 4D from Monocular Egocentric Video(ECCV2026) ReViV reconstructs viewer-centric human motion (body, hand, and gaze) and view-centric scene geometry (camera and depth) from a single egocentric RGB video in a unified feed-forward model. For visualization, the initial camera pose and camera intrinsics are estimated with GeoCalib.
0:54

ReViV: Reconstructing the Viewer and the View in 4D from Monocular Egocentric Video(ECCV2026) ReViV reconstructs viewer-centric human motion (body, hand, and gaze) and view-centric scene geometry (camera and depth) from a single egocentric RGB video in a unified feed-forward model. For visualization, the initial camera pose and camera intrinsics are estimated with GeoCalib.

Ryohei Sasaki@engineer

15,096 views • 2 months ago

EgoExoMoCap: Distributed Ego-Exo Human Motion Capture ECCV 2026 (Oral) Meta Reality Labs/ETH Zürich With two or more people wearing smart glasses, EgoExoMoCap combines each wearer’s egocentric (Ego) and exocentric (Exo) views to estimate full-body motion. Without bulky multi-camera setups or mocap suits, it reconstructs 3D human motion even under occlusions using head/wrist tracking and DINOv3-based visual features.
0:20

EgoExoMoCap: Distributed Ego-Exo Human Motion Capture ECCV 2026 (Oral) Meta Reality Labs/ETH Zürich With two or more people wearing smart glasses, EgoExoMoCap combines each wearer’s egocentric (Ego) and exocentric (Exo) views to estimate full-body motion. Without bulky multi-camera setups or mocap suits, it reconstructs 3D human motion even under occlusions using head/wrist tracking and DINOv3-based visual features.

Ryohei Sasaki@engineer

26,236 views • 2 months ago

MotionGPT: Human Motion as a Foreign Language paper page: Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this "motion vocabulary", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.
0:48

MotionGPT: Human Motion as a Foreign Language paper page: Though the advancement of pre-trained large language models unfolds, the exploration of building a unified model for language and other multi-modal data, such as motion, remains challenging and untouched so far. Fortunately, human motion displays a semantic coupling akin to human language, often perceived as a form of body language. By fusing language data with large-scale motion models, motion-language pre-training that can enhance the performance of motion-related tasks becomes feasible. Driven by this insight, we propose MotionGPT, a unified, versatile, and user-friendly motion-language model to handle multiple motion-relevant tasks. Specifically, we employ the discrete vector quantization for human motion and transfer 3D motion into motion tokens, similar to the generation process of word tokens. Building upon this "motion vocabulary", we perform language modeling on both motion and text in a unified manner, treating human motion as a specific language. Moreover, inspired by prompt learning, we pre-train MotionGPT with a mixture of motion-language data and fine-tune it on prompt-based question-and-answer tasks. Extensive experiments demonstrate that MotionGPT achieves state-of-the-art performances on multiple motion tasks including text-driven motion generation, motion captioning, motion prediction, and motion in-between.

AK

125,403 views • 3 years ago

Champ Controllable and Consistent Human Image Animation with 3D Parametric Guidance In this study, we introduce a methodology for human image animation by leveraging a 3D human parametric model within a latent diffusion framework to enhance shape alignment and motion
2:10

Champ Controllable and Consistent Human Image Animation with 3D Parametric Guidance In this study, we introduce a methodology for human image animation by leveraging a 3D human parametric model within a latent diffusion framework to enhance shape alignment and motion

AK

194,605 views • 2 years ago

Releasing EgoSim, a simulator for body-worn cameras. #NeurIPS2024 D&B EgoSim takes real mocap data (e.g., AMASS) and synthesizes multi-modal egocentric videos Plus: MultiEgoView, a real dataset from 6 GoPro cameras and ground-truth 3D poses during several activities (13 people)
0:45

Releasing EgoSim, a simulator for body-worn cameras. #NeurIPS2024 D&B EgoSim takes real mocap data (e.g., AMASS) and synthesizes multi-modal egocentric videos Plus: MultiEgoView, a real dataset from 6 GoPro cameras and ground-truth 3D poses during several activities (13 people)

Christian Holz

24,420 views • 1 year ago

Introducing Gen 2 spatial motion: 3D motion capture, full-body dynamics, joint torques, ground reaction forces, advanced motion retargeting, motion prediction. Works on AI video, phones, cinema cams, stadia. Launched at Movement Day, BAFTA London.
1:28

Introducing Gen 2 spatial motion: 3D motion capture, full-body dynamics, joint torques, ground reaction forces, advanced motion retargeting, motion prediction. Works on AI video, phones, cinema cams, stadia. Launched at Movement Day, BAFTA London.

Move AI

68,784 views • 1 year ago

byebye expensive motion tracking equipment 👋 ai makes motion capturing so easy now! nvidia presented GENMO last week, a new model that can generate and estimate human motion from text, audio, video, and 3D keyframes
0:26

byebye expensive motion tracking equipment 👋 ai makes motion capturing so easy now! nvidia presented GENMO last week, a new model that can generate and estimate human motion from text, audio, video, and 3D keyframes

Dreaming Tulpa 🥓👑

73,693 views • 1 year ago

We are publishing our first deep dive on what we believe is one of the most challenging layers in egocentric data - SLAM and VIO in the context of long-horizon state tracking. We break down how SLAM and VIO fail in egocentric settings - visual features vanish at close range, depth sensors saturate, fast head motion blurs frames, and these failures don't always occur in isolation. They hit at the exact same moment, leading to compounding errors and making the downstream data unusable. We believe the foundation for high-quality egocentric data demands sub-centimeter precision over long episodes ranging from a few minutes to up to an hour.
0:33

We are publishing our first deep dive on what we believe is one of the most challenging layers in egocentric data - SLAM and VIO in the context of long-horizon state tracking. We break down how SLAM and VIO fail in egocentric settings - visual features vanish at close range, depth sensors saturate, fast head motion blurs frames, and these failures don't always occur in isolation. They hit at the exact same moment, leading to compounding errors and making the downstream data unusable. We believe the foundation for high-quality egocentric data demands sub-centimeter precision over long episodes ranging from a few minutes to up to an hour.

FPV Labs

42,211 views • 6 months ago

Introducing 👀Stereo4D👀 A method for mining 4D from internet stereo videos. It enables large-scale, high-quality, dynamic, *metric* 3D reconstructions, with camera poses and long-term 3D motion trajectories. We used Stereo4D to make a dataset of over 100k real-world 4D scenes.
1:04

Introducing 👀Stereo4D👀 A method for mining 4D from internet stereo videos. It enables large-scale, high-quality, dynamic, *metric* 3D reconstructions, with camera poses and long-term 3D motion trajectories. We used Stereo4D to make a dataset of over 100k real-world 4D scenes.

Linyi Jin

94,780 views • 1 year ago

introducing openego the largest opensource POV manipulation dataset 1107 hours, 120M frames, 290 tasks, 600+ environments task, language actions & 3D hand pose annotations we are collecting egocentric data at scale using AI glasses dm me if your interested in egocentric data
0:23

introducing openego the largest opensource POV manipulation dataset 1107 hours, 120M frames, 290 tasks, 600+ environments task, language actions & 3D hand pose annotations we are collecting egocentric data at scale using AI glasses dm me if your interested in egocentric data

ahad

44,008 views • 1 year ago

📢Animating the Uncaptured 📢 We animate 3D humanoid meshes using video diffusion priors given a text prompt. 🎥 🌍 Realistic motion generation for 3D characters - without motion capture! 🚀 Great work by Marc Benedí Angela Dai
0:24

📢Animating the Uncaptured 📢 We animate 3D humanoid meshes using video diffusion priors given a text prompt. 🎥 🌍 Realistic motion generation for 3D characters - without motion capture! 🚀 Great work by Marc Benedí Angela Dai

Matthias Niessner

11,716 views • 1 year ago