ssan's Introduction

Spectral-Spatial Attention Network for Hyperspectral Image Classification

Paper

Spectral-Spatial Attention Networks for Hyperspectral Image Classification

Please cite our paper if you find it useful for your research.

@Article{ssan,
  AUTHOR = {Mei, Xiaoguang and Pan, Erting and Ma, Yong and Dai, Xiaobing and Huang, Jun  and Fan, Fan and Du, Qinglei and Zheng, Hong and Ma, Jiayi},
  TITLE = {Spectral-Spatial Attention Networks for Hyperspectral Image Classification},
  JOURNAL = {Remote Sensing},
  VOLUME = {11},
  YEAR = {2019},
  NUMBER = {8},
  ARTICLE-NUMBER = {963},
  URL = {http://www.mdpi.com/2072-4292/11/8/963},
  ISSN = {2072-4292},
  DOI = {10.3390/rs11080963}
}

Installation

Install Tensorflow 1.9.0 with Python 3.6.
Clone this repo

git clone https://github.com/EtPan/SSAN

Dataset

Download the Pavia Center dataset and its corresponding ground-truth map.

Usage

1. Change the file path

Replace the file path for the hyperspectral data in save_indices.py and indices.py;Replace the file path for the ckpt files of the model in ssan.py and ssan_test.py.

2. Split the dataset

Run save_indices.py .

3. Training

Run ssan.py.

4. Testing and Evaluation

Run ssan_test.py.

ssan's People

Stargazers

ssan's Issues

The question on the OA

Thank you sharing the code.
But after running the project with the same Parameter Setting, I just gain the OA about 0.94 in the test of Pavia Center Dataset. Could you tell me how to reach the same OA like you or maybe I have something wrong with the parameter settings?
Thank you very much