A Whale’s Tail - Finding the Right Whale in an Uncertain World
Abstract Explainable machine learning and uncertainty quantification have emerged as promising approaches to check the suitability and understand the decision process of a data-driven model, to learn new insights from data, but also to get more information about the quality of a specific observation...
Main Authors: | , , , , |
---|---|
Format: | Book Part |
Language: | unknown |
Published: |
Springer International Publishing
2022
|
Subjects: | |
Online Access: | http://dx.doi.org/10.1007/978-3-031-04083-2_15 https://link.springer.com/content/pdf/10.1007/978-3-031-04083-2_15 |
id |
crspringernat:10.1007/978-3-031-04083-2_15 |
---|---|
record_format |
openpolar |
spelling |
crspringernat:10.1007/978-3-031-04083-2_15 2024-03-10T08:35:11+00:00 A Whale’s Tail - Finding the Right Whale in an Uncertain World Marcos, Diego Kierdorf, Jana Cheeseman, Ted Tuia, Devis Roscher, Ribana 2022 http://dx.doi.org/10.1007/978-3-031-04083-2_15 https://link.springer.com/content/pdf/10.1007/978-3-031-04083-2_15 unknown Springer International Publishing https://creativecommons.org/licenses/by/4.0 https://creativecommons.org/licenses/by/4.0 xxAI - Beyond Explainable AI Lecture Notes in Computer Science page 297-313 ISSN 0302-9743 1611-3349 ISBN 9783031040825 9783031040832 book-chapter 2022 crspringernat https://doi.org/10.1007/978-3-031-04083-2_15 2024-02-13T17:15:07Z Abstract Explainable machine learning and uncertainty quantification have emerged as promising approaches to check the suitability and understand the decision process of a data-driven model, to learn new insights from data, but also to get more information about the quality of a specific observation. In particular, heatmapping techniques that indicate the sensitivity of image regions are routinely used in image analysis and interpretation. In this paper, we consider a landmark-based approach to generate heatmaps that help derive sensitivity and uncertainty information for an application in marine science to support the monitoring of whales. Single whale identification is important to monitor the migration of whales, to avoid double counting of individuals and to reach more accurate population estimates. Here, we specifically explore the use of fluke landmarks learned as attention maps for local feature extraction and without other supervision than the whale IDs. These individual fluke landmarks are then used jointly to predict the whale ID. With this model, we use several techniques to estimate the sensitivity and uncertainty as a function of the consensus level and stability of localisation among the landmarks. For our experiments, we use images of humpback whale flukes provided by the Kaggle Challenge “Humpback Whale Identification” and compare our results to those of a whale expert. Book Part Humpback Whale Springer Nature 297 313 |
institution |
Open Polar |
collection |
Springer Nature |
op_collection_id |
crspringernat |
language |
unknown |
description |
Abstract Explainable machine learning and uncertainty quantification have emerged as promising approaches to check the suitability and understand the decision process of a data-driven model, to learn new insights from data, but also to get more information about the quality of a specific observation. In particular, heatmapping techniques that indicate the sensitivity of image regions are routinely used in image analysis and interpretation. In this paper, we consider a landmark-based approach to generate heatmaps that help derive sensitivity and uncertainty information for an application in marine science to support the monitoring of whales. Single whale identification is important to monitor the migration of whales, to avoid double counting of individuals and to reach more accurate population estimates. Here, we specifically explore the use of fluke landmarks learned as attention maps for local feature extraction and without other supervision than the whale IDs. These individual fluke landmarks are then used jointly to predict the whale ID. With this model, we use several techniques to estimate the sensitivity and uncertainty as a function of the consensus level and stability of localisation among the landmarks. For our experiments, we use images of humpback whale flukes provided by the Kaggle Challenge “Humpback Whale Identification” and compare our results to those of a whale expert. |
format |
Book Part |
author |
Marcos, Diego Kierdorf, Jana Cheeseman, Ted Tuia, Devis Roscher, Ribana |
spellingShingle |
Marcos, Diego Kierdorf, Jana Cheeseman, Ted Tuia, Devis Roscher, Ribana A Whale’s Tail - Finding the Right Whale in an Uncertain World |
author_facet |
Marcos, Diego Kierdorf, Jana Cheeseman, Ted Tuia, Devis Roscher, Ribana |
author_sort |
Marcos, Diego |
title |
A Whale’s Tail - Finding the Right Whale in an Uncertain World |
title_short |
A Whale’s Tail - Finding the Right Whale in an Uncertain World |
title_full |
A Whale’s Tail - Finding the Right Whale in an Uncertain World |
title_fullStr |
A Whale’s Tail - Finding the Right Whale in an Uncertain World |
title_full_unstemmed |
A Whale’s Tail - Finding the Right Whale in an Uncertain World |
title_sort |
whale’s tail - finding the right whale in an uncertain world |
publisher |
Springer International Publishing |
publishDate |
2022 |
url |
http://dx.doi.org/10.1007/978-3-031-04083-2_15 https://link.springer.com/content/pdf/10.1007/978-3-031-04083-2_15 |
genre |
Humpback Whale |
genre_facet |
Humpback Whale |
op_source |
xxAI - Beyond Explainable AI Lecture Notes in Computer Science page 297-313 ISSN 0302-9743 1611-3349 ISBN 9783031040825 9783031040832 |
op_rights |
https://creativecommons.org/licenses/by/4.0 https://creativecommons.org/licenses/by/4.0 |
op_doi |
https://doi.org/10.1007/978-3-031-04083-2_15 |
container_start_page |
297 |
op_container_end_page |
313 |
_version_ |
1793131493048975360 |