Abstract
Systems and methods are described that can generate a predictive, first-person visual preview of a navigation route. A technique can involve conditioning a generative video synthesis framework with structured data, such as geometric and semantic information. This data, which may include three-dimensional pose graphs, depth information, and building footprints, can be used to constrain the synthesis to a physically plausible path. This approach can produce a topologically consistent preview that may be synchronized to a user's pace on a computing device (e.g., a smartphone, smart watch, or augmented reality glasses). The system can also render current conditions, like weather, and visually emphasize navigational landmarks to provide a contextually relevant guide for wayfinding.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Shin, D, "Geometry-Conditioned Synthesis of First-Person Visual Navigation Previews", Technical Disclosure Commons, (October 21, 2026)
https://www.tdcommons.org/dpubs_series/12365