/FSSF-Net-for-HSI-Classification

FSSF-Net for HSI Classification designed in a factorized manner

Primary LanguagePython

FSSF-Net-for-HSI-Classification

FSSF-Net for HSI Classification designed in a factorized manner.

avatar

Paper

Pre-print version

This work has been accepted by IEEE Access titled "Deep Neural Network Based Hyperspectral Pixel Classification With Factorized Spectral-Spatial Feature Representation". Please search for it for more details.

Code

  1. run.py: set the parameters and train the model.
  2. data.py: used for data process.
  3. evaluate.py: capsulate functions used for evaluation metrics such as OA, AA, Kappa.
  4. net.py: define all sorts of net's architectures.

Dataset

Please visit Hyperspectral Remote Sensing Scenes.

Requirements

Keras 2.0.8

Tensorflow 1.4.0

Compatibility