This is a PyTorch implementation of a rotation equivariant U-Net model for segmentation and classification of deforestation areas as described in our paper:
https://arxiv.org/abs/2110.13097
conda create --name roteqseg python=3.8
conda activate roteqseg
python setup.py install
The dataset used in this project is from https://stanfordmlgroup.github.io/projects/forestnet/ and can be dowloaded from their page.
Download the dataset into the roteqseg folder and unzip:
unzip ForestNetDataset.zip
We provide both a non-rotation equivariant model and a rotation equivariant model to allow a comparison to be made between the two models.
cd roteqseg
mkdir Outputs
python trainer.py --savedir 'code_test' --epochs 12 --model 'unet'
python trainer.py --savedir 'eq_code_test' -- epochs 12 --model 'unet_eq'
We also provide a notebook file, Post_Run_Analysis.ipynb, which performs the post processing we used to generate figures and numeric results used in the paper.
Please cite our paper if you make use of our work:
@article{mitton2021rotation,
title={Rotation Equivariant Deforestation Segmentation and Driver Classification},
author={Mitton, Joshua and Murray-Smith, Roderick},
journal={NeurIPS 2021 Workshop on Tackling Climate Change with Machine Learning},
year={2021}
}