Gaussian-SLAM: Photo-realistic Dense SLAM with Gaussian Splatting
This addresses the problem of real-time, high-quality 3D reconstruction and rendering for robotics or AR/VR applications, representing an incremental improvement over prior neural SLAM methods.
The paper tackles dense SLAM by using 3D Gaussians for scene representation, achieving interactive-time reconstruction and photo-realistic rendering from single-camera RGBD videos with competitive or superior performance in mapping, tracking, and rendering compared to existing neural dense SLAM methods.
We present a dense simultaneous localization and mapping (SLAM) method that uses 3D Gaussians as a scene representation. Our approach enables interactive-time reconstruction and photo-realistic rendering from real-world single-camera RGBD videos. To this end, we propose a novel effective strategy for seeding new Gaussians for newly explored areas and their effective online optimization that is independent of the scene size and thus scalable to larger scenes. This is achieved by organizing the scene into sub-maps which are independently optimized and do not need to be kept in memory. We further accomplish frame-to-model camera tracking by minimizing photometric and geometric losses between the input and rendered frames. The Gaussian representation allows for high-quality photo-realistic real-time rendering of real-world scenes. Evaluation on synthetic and real-world datasets demonstrates competitive or superior performance in mapping, tracking, and rendering compared to existing neural dense SLAM methods.