The task of roof damage classification and segmentation from overhead imagery presents unique challenges. In this work we choose to address the challenge posed due to strong class imbalance. We propose four distinct techniques that aim at mitigating this problem. Through a new scheme that feeds the data to the network by oversampling the minority classes, and three other network architectural improvements, we manage to boost the macro-averaged F1-score of a model by 39.9 percentage points, thus achieving improved segmentation performance, especially on the minority classes.
@article{arxiv.2010.07151,
title = {Multi-class segmentation under severe class imbalance: A case study in roof damage assessment},
author = {Jean-Baptiste Boin and Nat Roth and Jigar Doshi and Pablo Llueca and Nicolas Borensztein},
journal= {arXiv preprint arXiv:2010.07151},
year = {2020}
}
Comments
Submitted to the Artificial Intelligence for Humanitarian Assistance and Disaster Response Workshop at NeurIPS 2020