Abstract
Watching two-dimensional (2D) media in extended reality can involve static virtual environments that may not be contextually linked to the content being viewed, which can reduce immersion. Systems and techniques are described for dynamically generating a 360-degree virtual environment from a 2D media stream. A system may analyze a buffer of video frames using computer vision methods to extract scene context, including three-dimensional geometry and lighting information. This extracted data can then be used as a seed for a generative artificial intelligence model to perform spatial outpainting, extending the original 2D scene into a volumetric environment. This process can transform a media viewing session into an interactive and spatially aware experience, generating a surrounding world that can be contextually matched to the content and support six-degrees-of-freedom movement.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Daftari, Dev and Verma, Akash, "Dynamically Generating Immersive Environments for Extended Reality From 2D Media", Technical Disclosure Commons, (August 18, 2026)
https://www.tdcommons.org/dpubs_series/11409