MSG-Loc

[IEEE RA-L'26] This repository is the official implementation of "MSG-Loc: Multi-Label Likelihood-based Semantic Graph Matching for Object-Level Global Localization".
Gihyeon Lee, Jungwoo Lee, Juwon Kim, Young-Sik Shin^†, Younggun Cho^†

Spatial AI and Robotics Lab (SPARO)  
  Korea Institute of Machinery & Materials (KIMM) & Kyungpook National University (KNU)

---

NEWS

• [Dec, 2025] Project page is now available.
• [Nov, 2025] 🎉 MSG-Loc has been accepted by IEEE Robotics and Automation Letters (RA-L).

---

Citation

If you find this repository useful, please consider citing:

@ARTICLE{lee2026msgloc,
  author={Lee, Gihyeon and Lee, Jungwoo and Kim, Juwon and Shin, Young-Sik and Cho, Younggun},
  journal={IEEE Robotics and Automation Letters}, 
  title={MSG-Loc: Multi-Label Likelihood-Based Semantic Graph Matching for Object-Level Global Localization}, 
  year={2026},
  volume={11},
  number={2},
  pages={2066-2073},
  keywords={Semantics;Location awareness;Simultaneous localization and mapping;Uncertainty;Three-dimensional displays;Artificial intelligence;Object oriented modeling;Nearest neighbor methods;Pose estimation;Maximum likelihood estimation;Semantic scene understanding;localization;graph matching;object-based SLAM},
  doi={10.1109/LRA.2025.3643293}
}

---

Contact

• Gihyeon Lee (leekh951@inha.edu)

---