![[2602.15181] Time-Archival Camera Virtualization for Sports and Visual Performances](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.15181] Time-Archival Camera Virtualization for Sports and Visual Performances
Summary
This paper presents a novel approach to camera virtualization for sports and visual performances, enabling photorealistic rendering from multiple camera views and efficient time-archival capabilities.
Why It Matters
The advancement of camera virtualization has significant implications for sports broadcasting and live performances, allowing for enhanced viewer experiences through retrospective rendering and analysis. This research addresses existing limitations in dynamic scene rendering, making it relevant for industries reliant on real-time visual content.
Key Takeaways
- Introduces a neural volume rendering approach for camera virtualization.
- Supports time-archival, allowing users to revisit past dynamic scenes.
- Addresses challenges in rendering dynamic scenes with multiple subjects.
- Enhances visual quality through neural representation learning.
- Aims to improve sports broadcasting and live event analysis.
Computer Science > Computer Vision and Pattern Recognition arXiv:2602.15181 (cs) [Submitted on 16 Feb 2026] Title:Time-Archival Camera Virtualization for Sports and Visual Performances Authors:Yunxiao Zhang, William Stone, Suryansh Kumar View a PDF of the paper titled Time-Archival Camera Virtualization for Sports and Visual Performances, by Yunxiao Zhang and 2 other authors View PDF HTML (experimental) Abstract:Camera virtualization -- an emerging solution to novel view synthesis -- holds transformative potential for visual entertainment, live performances, and sports broadcasting by enabling the generation of photorealistic images from novel viewpoints using images from a limited set of calibrated multiple static physical cameras. Despite recent advances, achieving spatially and temporally coherent and photorealistic rendering of dynamic scenes with efficient time-archival capabilities, particularly in fast-paced sports and stage performances, remains challenging for existing approaches. Recent methods based on 3D Gaussian Splatting (3DGS) for dynamic scenes could offer real-time view-synthesis results. Yet, they are hindered by their dependence on accurate 3D point clouds from the structure-from-motion method and their inability to handle large, non-rigid, rapid motions of different subjects (e.g., flips, jumps, articulations, sudden player-to-player transitions). Moreover, independent motions of multiple subjects can break the Gaussian-tracking assumptions commonly use...