GaussianFlow SLAM: Monocular Gaussian Splatting SLAM Guided by GaussianFlow

2026-04-16

AI Summary

Monocular Gaussian splatting SLAM suffers from structural degeneracy and pose drift due to the lack of reliable geometric constraints. This work addresses this limitation by introducing optical flow as a geometry-aware guidance signal and proposes GaussianFlow alignment constraints that jointly optimize Gaussian projection motion and optical flow to simultaneously enhance scene geometry reconstruction and camera pose estimation accuracy. Furthermore, an adaptive Gaussian densification and pruning mechanism based on normalized reprojection error is devised to dynamically refine the Gaussian distribution during optimization. Evaluated on public benchmarks, the proposed method significantly outperforms existing approaches, achieving state-of-the-art performance in both rendering fidelity and tracking precision.

Technology Category

• Computer Vision: Motion & Tracking
• Intelligent Robots: State Estimation
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Gaussian splatting has recently gained traction as a compelling map representation for SLAM systems, enabling dense and photo-realistic scene modeling. However, its application to monocular SLAM remains challenging due to the lack of reliable geometric cues from monocular input. Without geometric supervision, mapping or tracking could fall in local-minima, resulting in structural degeneracies and inaccuracies. To address this challenge, we propose GaussianFlow SLAM, a monocular 3DGS-SLAM that leverages optical flow as a geometry-aware cue to guide the optimization of both the scene structure and camera poses. By encouraging the projected motion of Gaussians, termed GaussianFlow, to align with the optical flow, our method introduces consistent structural cues to regularize both map reconstruction and pose estimation. Furthermore, we introduce normalized error-based densification and pruning modules to refine inactive and unstable Gaussians, thereby contributing to improved map quality and pose accuracy. Experiments conducted on public datasets demonstrate that our method achieves superior rendering quality and tracking accuracy compared with state-of-the-art algorithms. The source code is available at: https://github.com/url-kaist/gaussianflow-slam.

Problem

• Monocular SLAM
• Gaussian Splatting
• Geometric Cues
• Structural Degeneracy
• Optical Flow

Innovation

• Gaussian Splatting
• Optical Flow
• Monocular SLAM
• Geometry-aware Regularization
• 3D Reconstruction