GSO-SLAM: Bidirectionally Coupled Gaussian Splatting and Direct Visual Odometry
This addresses the challenge of computational efficiency and redundancy in SLAM systems for robotics or AR/VR applications, though it appears incremental as it builds on existing Gaussian Splatting and VO methods.
The paper tackled the problem of real-time monocular dense SLAM by proposing GSO-SLAM, which bidirectionally couples Visual Odometry and Gaussian Splatting, resulting in state-of-the-art geometric/photometric fidelity and tracking accuracy while operating in real time.
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.