DOCSEARCHbyAlgolia
Submit Docs

MonoGS (Gaussian Splatting SLAM)

repository·main·Indexed 23 days ago

https://github.com/muskie82/monogs

A dense SLAM system presented at CVPR 2024 that utilizes 3D Gaussian Splatting for dense reconstruction and tracking. It supports monocular, RGB-D, and stereo inputs.

Tokens
901
Snippets
2
Records
4
Agent score
35%

What's inside MonoGS

  1. Understand the Gaussian-Splatting License terms

    main

    The gaussian-splatting software is owned by Inria and the Max Planck Institut für Informatik (MPII). The license is designed primarily for the research community (both academic and industrial) to use, test, and evaluate the software.

    Permitted Use

    • Non-commercial use only: The software may be used for research and/or evaluation purposes.
    • Rights granted: You are granted a non-exclusive, royalty-free license to reproduce, prepare derivative works of, publicly display, publicly perform, and distribute your work (which includes the software and any additions/derivatives).

    Restrictions and Requirements

    • No Commercial Use: You cannot use, exploit, or distribute the software for commercial purposes without prior and explicit consent from the licensors. For unauthorized commercial use, contact stip-sophia.transfert@inria.fr.
    • Redistribution: If you redistribute the work, you must:
      1. Do so under this same License.
      2. Include a complete copy of this License.
      3. Retain all original copyright, patent, trademark, or attribution notices.
    • Derivative Works: You may apply different terms to your derivative works only if those terms maintain the non-commercial use limitation defined in this license and you clearly identify which works are subject to your terms.
    • Citations: When using the software for publications or results, users are strongly encouraged to cite the corresponding publications as specified in the software's documentation.

  2. Understand MonoGS third-party dependencies and licenses

    main

    MonoGS includes several third-party codebases and linked libraries. When using MonoGS, you must adhere to the specific licenses of these components:

    • 3D Gaussian Splatting: Code located in the gaussian_splatting folder. Follow the 3D Gaussian Splatting license.
    • Tiny Gaussian Splatting Viewer: Code located in the gui/gl_render folder. Follow the Tiny Gaussian Splatting Viewer license.
    • Submodules: Any code located in the submodules folder is subject to the individual licenses of each respective repository.

  3. Install MonoGS

    main

    To install MonoGS, clone the repository recursively to include submodules and set up the environment using Conda.

    Note on Dependencies: Depending on your hardware, you may need to manually adjust the pytorch and cudatoolkit versions in environment.yml.

    Recommended test setups:

    • 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
    git clone https://github.com/muskie82/MonoGS.git --recursive
cd MonoGS
conda env create -f environment.yml
conda activate MonoGS

  4. Permissive license for utils/loss_utils.py

    main

    The file utils/loss_utils.py contains code based on the pytorch-ssim project by Evan Su, which is licensed under the MIT License. This specific file is subject to those permissive terms rather than the main Gaussian-Splatting license.

    Title: pytorch-ssim
Project code: https://github.com/Po-Hsun-Su/pytorch-ssim
Copyright Evan Su, 2017
License: https://github.com/Po-Hsun-Su/pytorch-ssim/blob/master/LICENSE.txt (MIT)