Towards Robust Contrail Detection by Mitigating Label Bias via a Probabilistic Deep Learning Model: A Preliminary Study

Yejun Lee, Eun-Kyeong Kim, Jaejun Yoo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Contrails, formed by jet flights, alter Earth's energy balance, prompting research into monitoring contrails and developing satellite-based automated contrail detection. This demand has advanced deep learning (DL)-based techniques. However, training DL algorithms to detect contrails has limitations: class imbalance, labeling difficulty, and a lack of reliable labeled datasets. We propose a probabilistic DL approach using P-UNet to alleviate label bias in contrail detection. By observing model outputs based on two labeled datasets, OpenContrails and MIT-Contrails, we found the probabilistic approach robust against potentially biased labels.
Original languageEnglish
Title of host publicationSIGSPATIAL '23: 31st ACM International Conference on Advances in Geographic Information Systems
Subtitle of host publicationHamburg, Germany, November 13-16, 2023
EditorsMaria Luisa Damiani, Matthias Renz, Ahmed Eldawy, Peer Kröger, Mario A. Nascimento
Place of PublicationNew-York
PublisherAssociation for Computing Machinery (ACM)
Pages1-2
Number of pages2
ISBN (Print)979-8-4007-0168-9
DOIs
Publication statusPublished - 22 Dec 2023

Keywords

  • Computing methodologies
  • Artificial intelligence
  • Mathematics of computing
  • Probability and statistics
  • Neural networks

Cite this