SGS-SLAM: Semantic Gaussian Splatting For Neural Dense SLAM

Abstract

We present SGS-SLAM, the first semantic visual SLAM system based on Gaussian Splatting. It incorporates appearance, geometry, and semantic features through multi-channel optimization, addressing the oversmoothing limitations of neural implicit SLAM systems in high-quality rendering, scene understanding, and object-level geometry. We introduce a unique semantic feature loss that effectively compensates for the shortcomings of traditional depth and color losses in object optimization. Through a semantic-guided keyframe selection strategy, we prevent erroneous reconstructions caused by cumulative errors. Extensive experiments demonstrate that SGS-SLAM delivers state-of-the-art performance in camera pose estimation, map reconstruction, precise semantic segmentation, and object-level geometric accuracy, while ensuring real-time rendering capabilities.

References in corpus (2)

• Instant Neural Graphics Primitives with a Multiresolution Hash Encoding
• Vox-Fusion: Dense Tracking and Mapping with Voxel-based Neural Implicit Representation

Cited by in corpus (7)

• NeSLAM: Neural Implicit Mapping and Self-Supervised Feature Tracking With Depth Completion and Denoising
• MG-SLAM: Structure Gaussian Splatting SLAM with Manhattan World Hypothesis
• Incremental Joint Learning of Depth, Pose and Implicit Scene Representation on Monocular Camera in Large-scale Scenes
• GSORB-SLAM: Gaussian Splatting SLAM benefits from ORB features and Transmittance information
• HI-SLAM2: Geometry-Aware Gaussian SLAM for Fast Monocular Scene Reconstruction
• DenseSplat: Densifying Gaussian Splatting SLAM with Neural Radiance Prior
• DynamicGSG: Dynamic 3D Gaussian Scene Graphs for Environment Adaptation