GSO-SLAM：双向耦合的高斯点云与直接视觉里程计
GSO-SLAM: Bidirectionally Coupled Gaussian Splatting and Direct Visual Odometry
AI总结：
GSO-SLAM通过双向耦合视觉里程计与高斯点云，实现实时单目密集SLAM，提升场景重建的几何和光度保真度。
摘要：
我们提出GSO-SLAM，一种实时单目密集SLAM系统，利用高斯场景表示。与现有方法将跟踪和映射统一到一个场景中，导致计算成本，或松散整合到结构良好的跟踪框架中，引入冗余不同，我们的方法双向耦合视觉里程计（VO）和高斯点云（GS）。具体而言，我们的方法在期望最大化（EM）框架内进行联合优化，使VO派生的半密集深度估计和GS表示同时得到精炼，而无需额外计算开销。此外，我们提出了高斯点云初始化，利用图像信息、关键帧姿态和像素关联从VO中产生接近最终高斯场景的近似，从而消除了对启发式方法的依赖。通过广泛的实验，我们验证了该方法的有效性，证明其不仅能够实时运行，还能实现重建场景的几何/光度保真度和跟踪精度的最先进水平。
英文摘要：
We propose GSO-SLAM, a real-time monocular dense SLAM system that leverages Gaussian scene representation. Unlike existing methods that couple tracking and mapping with a unified scene, incurring computational costs, or loosely integrate them with well-structured tracking frameworks, introducing redundancies, our method bidirectionally couples Visual Odometry (VO) and Gaussian Splatting (GS). Specifically, our approach formulates joint optimization within an Expectation-Maximization (EM) framework, enabling the simultaneous refinement of VO-derived semi-dense depth estimates and the GS representation without additional computational overhead. Moreover, we present Gaussian Splat Initialization, which utilizes image information, keyframe poses, and pixel associations from VO to produce close approximations to the final Gaussian scene, thereby eliminating the need for heuristic methods. Through extensive experiments, we validate the effectiveness of our method, showing that it not only operates in real time but also achieves state-of-the-art geometric/photometric fidelity of the reconstructed scene and tracking accuracy.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
