zoeyingz / cnn_hsic_mrf

This is a TensorFlow implementation of Convolutional Neural Network for Hyperspectral Image Classification

Hyperspectral Image Classification using Markov Random Field and a convolutional neural network

---

CNN-HSI-MRF

This is a TensorFlow implementation of Convolutional Neural Network (CNN) for the task of Hyperspectral Image Segmentation based on the code (https://github.com/KGPML/Hyperspectral), as described in our paper:

X. Cao, F. Zhou, L. Xu, D. Meng, Z. Xu and J. Paisley, Hyperspectral image segmentation with Markov random fields and a convolutional neural network, arXiv:1705.00727.

Another version of this code is in [https://github.com/xiangyongcao/CNN-MRF-v1]

Requirments

• tensorflow 0.12.1

How to use the code

• step 1: run the python code python cnn_train.py
• step 2: run the matlab code Demo_Post_MRF.m

Cite

This code is protected by the MIT licence. Please cite our paper if you use this code in your code:

@article{cao2017hyperspectral,
title={Hyperspectral image segmentation with Markov random fields and a convolutional neural network},
author={Cao, Xiangyong and Zhou, Feng and Xu, Lin and Meng, Deyu and Xu, Zongben and Paisley, John},
journal={arXiv preprint arXiv:1705.00727},
year={2017}
}

If you have any question in using the code, welcom to contact Xiangyong Cao ([email protected]).