delair-ai / DISCA

Contain the code for https://arxiv.org/abs/2009.11250

Presentation

This repository contains the code of DISCA from our paper: Interactive Learning for Semantic Segmentation in Earth Observation. In a nutshell, it consists in neural networks trained to perform semantic segmentation with human guidance. This builds on our previous work DISIR.

This repository is divided into two parts:

• train which contains the training code of the networks (README)
• qgs_plugin which contains the code of the QGIS plugin used to perform the interactive segmentation (README)

Install Python dependencies

conda create -n disca python=3.7 rtree gdal=2.4 opencv scipy shapely -c 'conda-forge' 
conda activate disca
pip install -r requirements.txt

To use

Please note that this repository has been tested on Ubuntu 18.4, QGIS 3.8 and python 3.7 only.

1. Download a segmentation dataset such as ISPRS Potsdam or INRIA dataset.
2. Prepare this dataset according to Dataset preprocessing in train/README.md.
3. Train a modelstill following train/README.md.
4. Install the QGIS plugin following qgs_plugin/README.md.
5. Follow How to start in qgs_plugin/README.md and start segmenting your data !

References

If you use this work for your projects, please take the time to cite our ECML-PKDD MACLEAN Workshop paper:

@inproceedings{lenczner2020interactive,
author = {Lenczner, G. and Chan-Hon-Tong, A. and Luminari, N. and Le Saux, B. and Le Besnerais, G.},
title = {Interactive Learning for Semantic Segmentation in Earth Observation},
booktitle = {ECML-PKDD MACLEAN Workshop},
year = {2020}
}

Licence

Code is released under the MIT license for non-commercial and research purposes only. For commercial purposes, please contact the authors.

See LICENSE for more details.

Authors

See AUTHORS.md

Acknowledgements

This work has been jointly conducted at Delair and ONERA-DTIS.