Glacier Calving Front Segmentation Using Attention U-Net
An essential climate variable to determine the tidewater glacier status is the location of the calving front position and the separation of seasonal variability from long-term trends. Previous studies have proposed deep learning-based methods to semi-automatically delineate the calving fronts of tid...
Main Authors: | , , , , , |
---|---|
Format: | Article in Journal/Newspaper |
Language: | unknown |
Published: |
arXiv
2021
|
Subjects: | |
Online Access: | https://dx.doi.org/10.48550/arxiv.2101.03247 https://arxiv.org/abs/2101.03247 |
id |
ftdatacite:10.48550/arxiv.2101.03247 |
---|---|
record_format |
openpolar |
spelling |
ftdatacite:10.48550/arxiv.2101.03247 2023-05-15T18:33:04+02:00 Glacier Calving Front Segmentation Using Attention U-Net Holzmann, Michael Davari, Amirabbas Seehaus, Thorsten Braun, Matthias Maier, Andreas Christlein, Vincent 2021 https://dx.doi.org/10.48550/arxiv.2101.03247 https://arxiv.org/abs/2101.03247 unknown arXiv Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 CC-BY Machine Learning cs.LG Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences Article CreativeWork article Preprint 2021 ftdatacite https://doi.org/10.48550/arxiv.2101.03247 2022-03-10T15:03:37Z An essential climate variable to determine the tidewater glacier status is the location of the calving front position and the separation of seasonal variability from long-term trends. Previous studies have proposed deep learning-based methods to semi-automatically delineate the calving fronts of tidewater glaciers. They used U-Net to segment the ice and non-ice regions and extracted the calving fronts in a post-processing step. In this work, we show a method to segment the glacier calving fronts from SAR images in an end-to-end fashion using Attention U-Net. The main objective is to investigate the attention mechanism in this application. Adding attention modules to the state-of-the-art U-Net network lets us analyze the learning process by extracting its attention maps. We use these maps as a tool to search for proper hyperparameters and loss functions in order to generate higher qualitative results. Our proposed attention U-Net performs comparably to the standard U-Net while providing additional insight into those regions on which the network learned to focus more. In the best case, the attention U-Net achieves a 1.5% better Dice score compared to the canonical U-Net with a glacier front line prediction certainty of up to 237.12 meters. Article in Journal/Newspaper Tidewater DataCite Metadata Store (German National Library of Science and Technology) |
institution |
Open Polar |
collection |
DataCite Metadata Store (German National Library of Science and Technology) |
op_collection_id |
ftdatacite |
language |
unknown |
topic |
Machine Learning cs.LG Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences |
spellingShingle |
Machine Learning cs.LG Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences Holzmann, Michael Davari, Amirabbas Seehaus, Thorsten Braun, Matthias Maier, Andreas Christlein, Vincent Glacier Calving Front Segmentation Using Attention U-Net |
topic_facet |
Machine Learning cs.LG Computer Vision and Pattern Recognition cs.CV FOS Computer and information sciences |
description |
An essential climate variable to determine the tidewater glacier status is the location of the calving front position and the separation of seasonal variability from long-term trends. Previous studies have proposed deep learning-based methods to semi-automatically delineate the calving fronts of tidewater glaciers. They used U-Net to segment the ice and non-ice regions and extracted the calving fronts in a post-processing step. In this work, we show a method to segment the glacier calving fronts from SAR images in an end-to-end fashion using Attention U-Net. The main objective is to investigate the attention mechanism in this application. Adding attention modules to the state-of-the-art U-Net network lets us analyze the learning process by extracting its attention maps. We use these maps as a tool to search for proper hyperparameters and loss functions in order to generate higher qualitative results. Our proposed attention U-Net performs comparably to the standard U-Net while providing additional insight into those regions on which the network learned to focus more. In the best case, the attention U-Net achieves a 1.5% better Dice score compared to the canonical U-Net with a glacier front line prediction certainty of up to 237.12 meters. |
format |
Article in Journal/Newspaper |
author |
Holzmann, Michael Davari, Amirabbas Seehaus, Thorsten Braun, Matthias Maier, Andreas Christlein, Vincent |
author_facet |
Holzmann, Michael Davari, Amirabbas Seehaus, Thorsten Braun, Matthias Maier, Andreas Christlein, Vincent |
author_sort |
Holzmann, Michael |
title |
Glacier Calving Front Segmentation Using Attention U-Net |
title_short |
Glacier Calving Front Segmentation Using Attention U-Net |
title_full |
Glacier Calving Front Segmentation Using Attention U-Net |
title_fullStr |
Glacier Calving Front Segmentation Using Attention U-Net |
title_full_unstemmed |
Glacier Calving Front Segmentation Using Attention U-Net |
title_sort |
glacier calving front segmentation using attention u-net |
publisher |
arXiv |
publishDate |
2021 |
url |
https://dx.doi.org/10.48550/arxiv.2101.03247 https://arxiv.org/abs/2101.03247 |
genre |
Tidewater |
genre_facet |
Tidewater |
op_rights |
Creative Commons Attribution 4.0 International https://creativecommons.org/licenses/by/4.0/legalcode cc-by-4.0 |
op_rightsnorm |
CC-BY |
op_doi |
https://doi.org/10.48550/arxiv.2101.03247 |
_version_ |
1766217350243155968 |