Revisiting Classical Bagging with Modern Transfer Learning for On-the-fly Disaster Damage Detector

TL;DR

This paper enhances disaster damage detection from aerial images by combining classical bagging with modern transfer learning, addressing data imbalance and improving detection accuracy in rapid assessment scenarios.

Contribution

It demonstrates the effectiveness of simple bagging combined with transfer learning for imbalanced disaster detection, outperforming recent advanced methods.

Findings

Significant performance improvement over existing methods.

Effective handling of data imbalance in disaster detection.

Outperforms recent disentanglement learning approaches.

Abstract

Automatic post-disaster damage detection using aerial imagery is crucial for quick assessment of damage caused by disaster and development of a recovery plan. The main problem preventing us from creating an applicable model in practice is that damaged (positive) examples we are trying to detect are much harder to obtain than undamaged (negative) examples, especially in short time. In this paper, we revisit the classical bootstrap aggregating approach in the context of modern transfer learning for data-efficient disaster damage detection. Unlike previous classical ensemble learning articles, our work points out the effectiveness of simple bagging in deep transfer learning that has been underestimated in the context of imbalanced classification. Benchmark results on the AIST Building Change Detection dataset show that our approach significantly outperforms existing methodologies, including the recently proposed disentanglement learning.