Kernel-SPD-Pooling

The code of PR 2019 paper "Learning a robust representation via a deep network on symmetric positive definite manifolds".

The caffe framework is required.

An example on the MINC dataset is put in

examples/minc

you can run it by

bash examples/minc/minc_conv_all.sh

If this code is helpful, we'd appreciate it if you could cite our paper

@article{gao2019learning,
  title={Learning a robust representation via a deep network on symmetric positive definite manifolds},
  author={Gao, Zhi and Wu, Yuwei and Bu, Xingyuan and Yu, Tan and Yuan, Junsong and Jia, Yunde},
  journal={Pattern Recognition},
  volume={92},
  pages={1--12},
  year={2019},
  publisher={Elsevier}
}