Graph Convolutional Subspace Clustering

This repository provides the implementations of EGCSC and EKGCSC model reported by "Y. Cai, Z. Zhang, Z. Cai, X. Liu, X. Jiang, and Q. Yan, “Graph convolutional subspace clustering: A robust subspace clustering framework for hyperspectral image,” IEEE Transactions on Geoscience and Remote Sensing, 2020"

If you would like to acknowledge our efforts, please cite the following paper:

@article{HSI-Clustering-GCSC-CAI-TGRS-2020,
title="Graph Convolutional Subspace Clustering: A Robust Subspace Clustering Framework for Hyperspectral Image",
author="Yaoming {Cai} and Zijia {Zhang} and Zhihua {Cai} and Xiaobo {Liu} and Xinwei {Jiang} and Qin {Yan}",
journal="IEEE Transactions on Geoscience and Remote Sensing",
note="doi: 10.1109/TGRS.2020.3018135",
year="2020",
}

Requirements

• Python >= 3.5

• Numpy <= 1.16.2

• Munkres

• SciPy

• Scikit-Learn

• Spectral Python (SPy)

Running

python demo.py

Dataset: HSI_Datasets/SalinasA_corrected.mat

original img shape: (83, 86, 204)

reduced img shape: (83, 86, 4)

x_patch tensor shape: (5348, 9, 9, 4)

final sample shape: (5348, 324), labels: [0. 1. 2. 3. 4. 5.]

============= EGCSC RESULTS =============

    OA      Kappa        NMI
0.9993     0.9971     0.9991

class accuracy: [1. 0.99702159 1. 1. 1. 1. ]

running time 42.296

============= EKGCSC RESULTS =============

    OA      Kappa        NMI
1.0000     1.0000     1.0000

class accuracy: [1. 1. 1. 1. 1. 1.]

running time 63.59

Reference hyper-parameter settings of EGCSC

=====  ===========  ===========  ===========
data    lambda            K          RO
=====  ===========  ===========  ===========
SaA      10             30          0.8
InP      100            30          0.4 (13*13 patch)
PaU      1000           20          0.6
===========================================

Reference hyper-parameter settings of EKGCSC

=====  ===========  ===========  ===========  ==========
data    lambda            K          RO         gamma
=====  ===========  ===========  ===========  ==========
SaA      100            30          0.8          0.2
InP      1e3            30          0.8          10 (13*13 patch)
PaU      6*1e4          30          0.8          100
========================================================