Research Analyzer
← Back ICRA 2026

MipSLAM: Alias-Free Gaussian Splatting SLAM

Yingzhao Li, Yan Li, Shixiong Tian, Yanjie Liu, lijun zhao, Gim Hee Lee

PDF

AI summary

Key figure (auto-extracted from paper)
MipSLAM eliminates aliasing artifacts and reduces trajectory drift in 3D Gaussian Splatting SLAM by introducing frequency-aware rendering and spectral pose optimization.
3D Gaussian Splatting Simultaneous Localization and Mapping Anti-aliasing Pose Graph Optimization Frequency-aware SLAM

Problem

Existing 3DGS-based SLAM systems produce severe aliasing artifacts and trajectory drift when camera parameters like resolution or focal length change, due to inadequate filtering and purely spatial optimization.

Approach

The framework replaces point-sampling with an Elliptical Adaptive Anti-aliasing algorithm that uses geometry-aware numerical integration, and reformulates pose graph optimization in the frequency domain to suppress high-frequency noise.

Key results

  • First frequency-aware 3DGS SLAM supporting arbitrary camera reconfiguration
  • Geometry-aware elliptical sampling for efficient anti-aliasing
  • Spectral pose graph optimization to suppress trajectory drift
  • State-of-the-art rendering and localization across multiple resolutions

Why it matters

Enables robust, high-fidelity 3D reconstruction and reliable robot navigation in dynamic environments with changing camera setups, benefiting autonomous robotics and AR/VR systems.

Abstract

This paper introduces MipSLAM, a frequency- aware 3D Gaussian Splatting (3DGS) SLAM framework ca- pable of high-fidelity anti-aliased novel view synthesis and robust pose estimation under varying camera configurations. Existing 3DGS-based SLAM systems often suffer from aliasing artifacts and trajectory drift due to inadequate filtering and purely spatial optimization. To overcome these limitations, we propose an Elliptical Adaptive Anti-aliasing (EAA) algorithm that approximates Gaussian contributions via geometry-aware numerical integration, avoiding costly analytic computation. Furthermore, we present a Spectral-Aware Pose Graph Op- timization (SA-PGO) module that reformulates trajectory esti- mation in the frequency domain, effectively suppressing high- frequency noise and drift through graph Laplacian analysis. Ex- tensive evaluations on Replica and TUM datasets demonstrate that MipSLAM achieves state-of-the-art rendering quality and localization accuracy across multiple resolutions.

Index terms

SLAM

Related papers