FlowAvatar runs entirely on a Meta Quest Pro, using only its headset and controller sensors to estimate SMPL-based full-body pose, hand articulation, facial expression, and eye gaze.
Consumer XR headsets can drive personalized full-body avatars in real time
FlowAvatar unifies body, hand, face, and eye tracking with look-alike appearance on a Meta Quest Pro, and user studies show the two factors interact to produce embodiment.
Big Tech
Chenfeng Gao · Taeyoung Yeon · Sungheon Park · Vasco Xu · Anish Prabhu · Mar Gonzalez-Franco · +1 more
Northwestern University · University of Chicago · Google
Research Digest··2 min read
Thread:3D Object Reconstruction
The authors present FlowAvatar, a system that reconstructs personalized look-alike avatars in real time on consumer XR hardware from sparse headset-mounted sensors.
Why this paper
From Google and 2 others · Released code · Part of 3D Object Reconstruction, now 7 papers
In one line
FlowAvatar creates real-time full-body avatars on consumer XR devices from sparse egocentric inputs.
What it released
Code
What we could check
- ✓Code link in the paper (github.com)
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ·No stated limitations found
- ·No benchmark numbers found
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§