Version 1.0, Copyright(c) December, 2025.

This is the source code of our AAAI paper, "Learning Topology-Driven Multi-Subspace Fusion for Grassmannian Deep Networks". This implementation is built upon the publicly available codes of SPDBN and GDLNet.. All the settings are consistent with the description in the main paper. If you have any queries, please do not hesitate to contact me 6243114042@stu.jiangnan.edu.cn.

Usage:

Step 1: Download the preprocessed FPHA dataset from the SymNet

Step 2: Unzip the dataset into '.\data'.

Step 3: Launch FPHA.py for a simple example.

@article{yu2025learning,
  title={Learning Topology-Driven Multi-Subspace Fusion for Grassmannian Deep Network},
  author={Yu, Xuan and Xu, Tianyang},
  journal={arXiv preprint arXiv:2511.08628},
  year={2025}
}

We would like to express our sincere gratitude to the following resources and contributors that greatly helped in the development of this project:

• Special thanks to SPDBN and GDLNet.