Note

May 31, 2026 · View on GitHub

MipSLAM: Alias-Free Gaussian Splatting SLAM

Yingzhao Li · Yan Li · Shixiong Tian · Yanjie Liu · Lijun Zhao* · Gim Hee Lee

(* Corresponding author)

This software implements the dense SLAM system presented in our paper MipSLAM: Alias-Free Gaussian Splatting SLAM at ICRA 2026.

Highlights

First frequency-aware 3DGS SLAM supporting arbitrary camera reconfiguration (intrinsics, resolution, zoom) with high-fidelity anti-aliasing.
Elliptical Adaptive Anti-aliasing (EAA): geometry-driven numerical integration that approximates analytic accuracy at a fraction of the cost.
Spectral-Aware Pose Graph Optimization (SA-PGO): models trajectories as spatiotemporal signals and leverages graph Laplacian spectral decomposition for drift-robust pose estimation.

Note

A version with higher localization accuracy and better rendering quality is coming soon. We will also provide rendering results at the original resolution on the Replica and TUM datasets. This allows readers to compare them with our results.

Getting Started

Installation

git clone https://github.com/yzli1998/MipSLAM.git --recursive
cd MipSLAM

Setup the environment.

conda env create -f environment.yml
conda activate MipSLAM

Depending on your setup, please change the dependency version of pytorch/cudatoolkit in environment.yml by following this document.

Our test setup were:

Ubuntu 20.04: pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cudatoolkit=11.6
Ubuntu 18.04: pytorch==1.12.1 torchvision==0.13.1 torchaudio==0.12.1 cudatoolkit=11.3

Downloading Datasets

Running the following scripts will automatically download datasets to the ./datasets folder.

TUM-RGBD dataset

bash scripts/download_tum.sh

Replica dataset

bash scripts/download_replica.sh

Run

Monocular

python slam.py --config configs/mono/tum/fr3_office.yaml

RGB-D

python slam.py --config configs/rgbd/tum/fr3_office.yaml

python slam.py --config configs/rgbd/replica/office0.yaml

Stereo (experimental)

python slam.py --config configs/stereo/euroc/mh02.yaml

Live demo with Realsense

First, you'll need to install pyrealsense2. Inside the conda environment, run:

pip install pyrealsense2

Connect the realsense camera to the PC on a USB-3 port and then run:

python slam.py --config configs/live/realsense.yaml

We tested the method with Intel Realsense d455. We recommend using a similar global shutter camera for robust camera tracking. Please avoid aggressive camera motion, especially before the initial BA is performed.

Evaluation

To evaluate our method, please add --eval to the command line argument:

python slam.py --config configs/mono/tum/fr3_office.yaml --eval

This flag will automatically run our system in a headless mode, and log the results including the rendering metrics.

Reproducibility

There might be minor differences between the released version and the results in the paper. Please bear in mind that multi-process performance has some randomness due to GPU utilisation. We run all our experiments on an RTX 4090, and the performance may differ when running with a different GPU.

Acknowledgement

This work incorporates many open-source codes. We extend our gratitude to the authors of the software.

Citation

If you found this code/work to be useful in your own research, please consider citing the following:

@inproceedings{li2026mipslam,
  title={{M}ip{SLAM}: {A}lias-{F}ree {G}aussian {S}platting {SLAM}},
  author={Yingzhao Li and Yan Li and Shixiong Tian and Yanjie Liu and Lijun Zhao and Gim Hee Lee},
  booktitle={Proceedings of the IEEE International Conference on Robotics and Automation (ICRA)},
  year={2026}
}