Holographic communication holds transformative potential for remote work, education, and telemedicine—but current systems require high bandwidth and expensive, specialised hardware. XRStream provides a solution: using AI to enable photorealistic holographic telepresence from simple 2D image inputs, dramatically lowering cost and infrastructure requirements. The system applies deep generative models and GPU acceleration at the edge to reconstruct and render realistic 3D avatars in real time, enabling immersive XR communication without 3D capture hardware.
The key objectives of XRStream are:
1.Real-time inference and model compression for ultra-fast performance.
2.GPU-accelerated rendering on affordable devices.
3.Low-latency communication over common high-bandwidth networks (e.g. 5G).
Target milestones include optimised inference models, GPU rendering deployment, end-to-end system integration, and full-scale user validation.The ultimate goal is to enable affordable holographic conferencing for broad market adoption.
Expected Impact
1.Sustainability and Performance: Delivers energy-efficient XR communications, reducing the carbon cost of high-quality remote interaction.(Based on SUSTAIN academic research)
2.Technology Accessibility: Makes immersive XR communication viable without specialised hardware, opening new markets in education, healthcare, and business.
3.Commercialisation: Ability to address a wide addressable marketing through licensing into a startup
