Autoencoders pretraining using clustering

DESCRIPTION:

Autoencoders pretraining using clustering.

v.1.0

Related to the work:

Improving Autoencoders Performance for Hyperspectral Unmixing using Clustering

Source code for the review process of the 14th Asian Conference on Intelligent Information and Database Systems (ACIIDS 2022).

LICENSE:

Copyright 2021 Institute of Theoretical and Applied Informatics, Polish Academy of Sciences (ITAI PAS) https://www.iitis.pl Authors:

• Bartosz Grabowski (ITAI PAS, ORCID ID: 0000−0002−2364−6547)
• Przemysław Głomb (ITAI PAS, ORCID ID: 0000−0002−0215−4674),
• Kamil Książek (ITAI PAS, ORCID ID: 0000−0002−0201−6220),
• Krisztián Buza (Sapientia Hungarian University of Transylvania, ORCID ID: 0000-0002-7111-6452)

This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by the Free Software Foundation, either version 3 of the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

You should have received a copy of the GNU General Public License along with this program. If not, see https://www.gnu.org/licenses/.

FUNCTIONALITY:

• Autoencoder training and evaluation for spectral unmixing task
• Autoencoder pretraining using clustering algorithm

FILES:

• ATE/*: Autoencoder Testing Environment files.
• cfg/*: Config files.
• grids/*.py: Files required to run comparison between baseline and pretraining-based autoencoder training.
• grids/run_exp.sh: File to run comparison between baseline and pretraining-based autoencoder training.
• grids/scripts/*: Simple scripts used for various purposes.
• grids/tests/*: Unit tests.

DATASETS:

All datasets have to be inserted into ./ATE/data/ folder.

USAGE:

Run the script using ./grids/run_exp.sh file. The script requires Samson and Jasper datasets in the ./ATE/data/ folder as well as saved models' weights in mpath (set by default to ./models).

To run demo version of the experiment, run ./grids/run_exp_demo.sh file. The results of the experiment will be generated in the ./results directory. Please note that this version of the script uses Custom dataset which is composed of random numbers, so the results too are going to be random.

DEPENDENCIES

The scripts are dependent on Autoencoder Testing Environment. Used datasets, as well as loaded models' weights follow the same structure.