Game-KFS: Game-Theory-Inspired Keyframe Selection for Hybrid Representation Visual SLAM
Shilang Chen, Bo Yang, Chaoqun Wang, Peidong Fang, Haifei Zhu, Weinan Chen, Yisheng Guan
AI summary
Problem
Current keyframe selection methods in hybrid VSLAM fail to simultaneously satisfy the high-precision tracking needs of discrete representations and the high-quality rendering requirements of field representations, often degrading one to optimize the other.
Approach
Game-KFS formulates keyframe selection as a dynamic trade-off between discrete geometric tracking and radiance field rendering, using an online-adapted weighted sum to balance these competing objectives in real time.
Key results
- A dynamically-weighted multi-objective framework for online keyframe selection
- Objective functions derived to balance discrete tracking accuracy and field rendering fidelity
- Significant improvements in tracking accuracy and scene reconstruction quality on public benchmarks and real-world tests
- Outperforms existing baselines (SplaTAM, Photo-SLAM, MonoGS) in both localization and rendering metrics
Why it matters
Enables autonomous robots and AR/VR systems to achieve both centimeter-level localization and photo-realistic dense mapping simultaneously.
Abstract
Hybrid representation Visual Simultaneous Local- ization and Mapping (VSLAM) systems combine the inherent strengths of both discrete and field representations. They promise high-precision tracking and photo-realistic dense mapping. How- ever, current keyframe selection methods in hybrid representation VSLAM struggle to satisfy both the high-precision tracking requirements of discrete representations and the high-quality rendering requirements of field representations. In this paper, we propose a game-theory-inspired keyframe selection approach that addresses the requirements of both representation types. We introduce two objective functions to comprehensively assess discrete point tracking and radiance field model rendering. By employing a game-theory-inspired framework, our method ef- fectively balances these objectives to achieve improved keyframe selection. Experimental results demonstrate that integrating our approach into a hybrid representation VSLAM system sig- nificantly enhances tracking accuracy and rendering quality, outperforming existing keyframe selection methods.