AI-empowered Pose Reconstruction for Real-time Synthesis of Remote Metaverse Avatars

The Metaverse, a synergistic blend of physical and virtual realities, is rapidly evolving as a hub for digital twins and digital avatars, offering transformative potential in domains ranging from smart manufacturing to kinematic examination. This paper introduces an AI-empowered framework for the synthesis of Metaverse avatars through a monocular camera-based system. Our approach integrates a shared control system and advanced multistage filtering for efine se nsor da ta, substantially enhancing pose precision and avatar realism. Experimental results demonstrate our framework's superiority in reducing data jitter and improving network transmission. By harnessing the power of AI and network optimization, our system ensures a cost-effective and accurate solution that enhances user interactivity and presence within the Metaverse. The implications of this work extend beyond immediate interaction benefits, setting a precedent for future immersive and inclusive Metaverse experiences. Future work will focus on expanding the framework's capacity for larger user interactions and incorporating richer features such as nuanced facial and gesture recognition to deliver fully expressive avatars, thereby elevating the Metaverse experience.