SG-Reg: Generalizable and Efficient
Scene Graph Registration

Submitted to IEEE T-RO
Chuhao Liu¹, Zhijian Qiao¹, Jieqi Shi^2,*, Ke Wang³, Peize Liu ¹ and Shaojie Shen¹

¹HKUST Aerial Robotics Group    ² NanJing University    ³Chang'an University   
^*Corresponding Author

News

• [Oct 2024] Paper submitted to IEEE T-RO.

We will publish the code once the paper is accepted.

To develop

• Scene graph network code and the trained network weight.
• Clean the data structure and upload the scene graph data for validation.
• Registration back-end in python interface.
• Remove unncessary dependencies.
• Validation the entire system in a new computer.
• A tutorial for running the validation.