Best practices to train deep models on imbalanced datasets—a case study on animal detection in aerial imagery

Benjamin Kellenberger*, Diego Marcos, Devis Tuia

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference paper

Abstract

We introduce recommendations to train a Convolutional Neural Network for grid-based detection on a dataset that has a substantial class imbalance. These include curriculum learning, hard negative mining, a special border class, and more. We evaluate the recommendations on the problem of animal detection in aerial images, where we obtain an increase in precision from 9% to 40% at high recalls, compared to state-of-the-art. Data related to this paper are available at: http://doi.org/10.5281/zenodo.609023.

Original languageEnglish
Title of host publicationMachine Learning and Knowledge Discovery in Databases - European Conference, ECML PKDD 2018, Proceedings
EditorsUlf Brefeld, Alice Marascu, Fabio Pinelli, Edward Curry, Brian MacNamee, Neil Hurley, Elizabeth Daly, Michele Berlingerio
PublisherSpringer Verlag
Pages630-634
ISBN (Print)9783030109967
DOIs
Publication statusPublished - 1 Jan 2019
EventEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML-PKDD 2018 - Dublin, Ireland
Duration: 10 Sep 201814 Sep 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11053 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

ConferenceEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML-PKDD 2018
CountryIreland
CityDublin
Period10/09/1814/09/18

Keywords

  • Class imbalance
  • Deep learning
  • Unmanned Aerial Vehicles

Fingerprint Dive into the research topics of 'Best practices to train deep models on imbalanced datasets—a case study on animal detection in aerial imagery'. Together they form a unique fingerprint.

Cite this