Deep Active Contour Models for Delineating Glacier Calving Fronts

Research output: Contribution to book/conference proceedings/anthology/reportConference contributionContributedpeer-review

Contributors

  • Konrad Heidler - , Technical University of Munich, German Aerospace Center (DLR) (e.V.) Location Oberpfaffenhofen (Author)
  • Lichao Mou - , Technical University of Munich, German Aerospace Center (DLR) (e.V.) Location Oberpfaffenhofen (Author)
  • Erik Loebel - , Chair of Geodetic Earth System Research (Author)
  • Mirko Scheinert - , Chair of Geodetic Earth System Research (Author)
  • Sébastien Lefèvre - , Université de Bretagne Sud (Author)
  • Xiao Xiang Zhu - , Technical University of Munich, German Aerospace Center (DLR) (e.V.) Location Oberpfaffenhofen (Author)

Abstract

We present a deep active contour model for detecting and delineating glacier calving fronts from satellite imagery. Contrary to existing deep learning-based calving front detectors, our model does not perform an intermediate segmentation or pixel-wise edge detection, but instead directly predicts the contour parametrized by a fixed number of vertices. The model works by first deriving feature maps from an input image, and then updating an initial contour in an iterative fashion. Evaluating on the CALFIN dataset, which maps calving fronts in Greenland, our model outperforms existing approaches. Code for the experiments and animated predictions can be found at https://github.com/khdlr/deep-acm

Details

Original languageEnglish
Title of host publicationIGARSS 2022 - 2022 IEEE International Geoscience and Remote Sensing Symposium
Pages4490-4493
Number of pages4
ISBN (electronic)9781665427920
Publication statusPublished - 28 Sept 2022
Peer-reviewedYes

External IDs

dblp conf/igarss/HeidlerMLSLZ22
Mendeley 399a5967-732f-3bdf-aed4-3b0948d03fcb
Scopus 85141895894
ORCID /0000-0002-0892-8941/work/142248916
ORCID /0000-0001-9874-9295/work/142255129

Keywords

DFG Classification of Subject Areas according to Review Boards

Subject groups, research areas, subject areas according to Destatis