arXiv cs.GRResearch
9 hours, 17 minutes ago • Deepti Parachuri, Chhayank Sahu, Sameer Singh Choudhary
A Browser-Based Gesture-Driven Avatar Interaction Framework for Metaverse Onboarding Environments
A new browser-deployable avatar interaction layer is aimed at metaverse onboarding spaces, where first-time users can explore virtual rooms and interact with embedded video, documents, and quizzes without picking up a controller. The system is built around real-time gesture recognition using Google MediaPipe, with hand, arm, and head input driving avatar actions directly in the browser.
Two locomotion approaches are bundled into the same framework: hand-raise navigation for more precise movement and in-place walking for a more physically immersive feel. That tradeoff matters for developers building onboarding or training experiences, because the interaction model can be tuned for either accuracy or embodiment depending on the task and the space.
The work was evaluated in a structured internal onboarding session with five participants. The practical takeaway is less about a single breakthrough feature and more about packaging controller-free interaction into a lightweight web stack that can be deployed broadly without requiring specialized hardware or a native client.
For teams building social VR, onboarding funnels, or browser-based spatial experiences, the interesting part is the integration problem: gesture recognition, avatar control, and content interaction all have to coexist cleanly in a low-friction web environment. The reported design trade-offs should be useful for anyone trying to make first-contact metaverse UX feel less like a demo and more like a usable product.
“users explore a set of virtual rooms as an avatar”
At a glance
what
A browser-based gesture-driven avatar interaction framework was built for metaverse onboarding environments.
who
Deepti Parachuri, Chhayank Sahu, and Sameer Singh Choudhary.
when
Submitted on 12 Aug 2026.
impact
Lets users navigate and interact in-browser using hand, arm, and head gestures instead of a controller-first interface.
SignalMixed
Promising UX gains, but only tested with five participants.
What do you think?
Start the discussion
Originalnavatar interactiononboardingmediapipe
Discuss
