📌 2026 Research Update. Between 2024 and 2026 this repo's pose-estimation work landed at the center of Physical AI — Vision-Language-Action (VLA) foundation models (OpenVLA, π0/π0.5, Gemini Robotics, NVIDIA GR00T N1, Figure Helix), world/physics simulators (Cosmos, Genesis, Newton), and humanoid platforms increasingly trained from human motion. See 2026 Physical AI & Embodied Intelligence for the sourced review and Mossland's revised angle: reposition the pose pipeline as the data + evaluation layer of embodied AI (not a metaverse side-quest).
The MosslandXR repository is dedicated to Extended Reality (XR) research, with a focus on developing and advancing technologies in Augmented Reality (AR), Virtual Reality (VR), and Mixed Reality (MR). This initiative is part of Mossland's effort to enhance the metaverse experience through innovative applications and immersive digital environments — and, increasingly, to feed the demonstration-data and evaluation needs of Physical AI.
- Innovation in XR: To explore and create cutting-edge XR technologies.
- Application Development: To develop practical applications within the Mossland ecosystem, enhancing user interaction and experience in the metaverse.
- Community Engagement: To engage with the community for feedback, collaboration, and shared learning.
Explore our extensive research into Extended Reality technologies and their applications in the Mossland ecosystem:
- XR Pose Estimation Research Overview - A comprehensive review of recent developments in pose estimation technologies for XR applications.
- RoboPoseGen Project Overview - Detailed exploration into synthetic pose data generation using Unreal Engine for robotic pose estimation.
- Humanoid Robot Pose Estimation Research - Research on pose estimation techniques tailored for humanoid robots.
For any inquiries, please reach out to us at contact@moss.land
- The Mossland team for their continuous support and innovation.
- The XR community for inspiring us with new ideas and perspectives.