Prototype framework for deep learning driven semantic segmentation of Arctic seabed imagery
The reliable mapping of the Arctic seabed is a prerequisite for marine spatial planning and environmental management. Underwater imagery (UWI) is one of the most common methods for mapping the seabed. The main advantage of this method is its simplicity, which enables rapid and cost-effective collect...
Published in: | Arctic Science |
---|---|
Main Authors: | , , , , |
Format: | Conference Object |
Language: | English |
Published: |
2021
|
Subjects: | |
Online Access: | https://vb.ktu.edu/KTU:ELABAPDB89723914&prefLang=en_US |
id |
ftkaunastuniv:oai:elaba:89723914 |
---|---|
record_format |
openpolar |
spelling |
ftkaunastuniv:oai:elaba:89723914 2023-05-15T14:21:50+02:00 Prototype framework for deep learning driven semantic segmentation of Arctic seabed imagery Buškus, Kazimieras Vaičiukynas, Evaldas Verikas, Antanas Medelytė, Saulė Olenin, Sergej 2021 application/pdf https://vb.ktu.edu/KTU:ELABAPDB89723914&prefLang=en_US eng eng info:eu-repo/semantics/altIdentifier/doi/10.1139/as-2021-0001 https://epubl.ktu.edu/object/elaba:89723914/89723914.pdf https://vb.ktu.edu/KTU:ELABAPDB89723914&prefLang=en_US info:eu-repo/semantics/openAccess Arctic science: Arctic Change 2020 conference book of abstracts = Compilation de Résumés pour la Conférence Arctic Change 2020: virtual event hosted by Université Laval, Québec, Canada, Ottawa, Ontario : Canadian science publishing, 2021, vol. 7, iss. 1, ID: 463, p. 85 ISSN 2368-7460 Long-Term Monitoring Data Management Underwater Imagery info:eu-repo/semantics/conferencePaper 2021 ftkaunastuniv https://doi.org/10.1139/as-2021-0001 2021-12-30T09:55:38Z The reliable mapping of the Arctic seabed is a prerequisite for marine spatial planning and environmental management. Underwater imagery (UWI) is one of the most common methods for mapping the seabed. The main advantage of this method is its simplicity, which enables rapid and cost-effective collection of large amounts of data. However, only a small part of information from UWI archives is being used due to labor-intensive and time-consuming effort, often requiring expert knowledge. Thanks to progress in deep learning (DL) methods, with convolutional neural networks demonstrating state-of-the-art in semantic segmentation tasks, new opportunities for automated large-scale analysis of seabed images have emerged. Although some tools exist for intermediate steps, multi-disciplinary teams still lack a unified framework, which could provide fast and reliable semantic segmentation and automate quantitative measurements for classes of interest. To address this issue, we present a prototype segmentation framework tested on the UWI imagery material collected from the upper subtidal zone of central Spitsbergen, Arctic. The system consists of two main layers: a user-friendly web front-end, based on industry leading React framework, and a back-end, relying on TensorFlow Serving API, responsible for managing and serving of DL models. A pilot experiment was conducted using two large 2D mosaics of a seabed transect, stitched from UWI video material, containing 3 classes — pebbles, tube dwelling polychaetes, and Ophiuroidea. Mosaic-based 2-fold cross-validation results revealed a high overlap between ground truth and predicted masks according to similarity coefficients — Dice of 0.86 and 0.84 as well as Jaccard of 0.77 and 0.72 for binary (Ophiuroidea detection) and multiclass segmentation tasks, respectively. Effective, accurate and fast segmentation of the seabed mosaic can be achieved if sufficient training annotations are available for the DL model. Further development is envisaged to enable tweaking capabilities for segmentation results as the post-processing phase. Conference Object Arctic Arctic Spitsbergen KTU ePubl (Kaunas University of Technology) Arctic Arctic Science 7 1 3 135 |
institution |
Open Polar |
collection |
KTU ePubl (Kaunas University of Technology) |
op_collection_id |
ftkaunastuniv |
language |
English |
topic |
Long-Term Monitoring Data Management Underwater Imagery |
spellingShingle |
Long-Term Monitoring Data Management Underwater Imagery Buškus, Kazimieras Vaičiukynas, Evaldas Verikas, Antanas Medelytė, Saulė Olenin, Sergej Prototype framework for deep learning driven semantic segmentation of Arctic seabed imagery |
topic_facet |
Long-Term Monitoring Data Management Underwater Imagery |
description |
The reliable mapping of the Arctic seabed is a prerequisite for marine spatial planning and environmental management. Underwater imagery (UWI) is one of the most common methods for mapping the seabed. The main advantage of this method is its simplicity, which enables rapid and cost-effective collection of large amounts of data. However, only a small part of information from UWI archives is being used due to labor-intensive and time-consuming effort, often requiring expert knowledge. Thanks to progress in deep learning (DL) methods, with convolutional neural networks demonstrating state-of-the-art in semantic segmentation tasks, new opportunities for automated large-scale analysis of seabed images have emerged. Although some tools exist for intermediate steps, multi-disciplinary teams still lack a unified framework, which could provide fast and reliable semantic segmentation and automate quantitative measurements for classes of interest. To address this issue, we present a prototype segmentation framework tested on the UWI imagery material collected from the upper subtidal zone of central Spitsbergen, Arctic. The system consists of two main layers: a user-friendly web front-end, based on industry leading React framework, and a back-end, relying on TensorFlow Serving API, responsible for managing and serving of DL models. A pilot experiment was conducted using two large 2D mosaics of a seabed transect, stitched from UWI video material, containing 3 classes — pebbles, tube dwelling polychaetes, and Ophiuroidea. Mosaic-based 2-fold cross-validation results revealed a high overlap between ground truth and predicted masks according to similarity coefficients — Dice of 0.86 and 0.84 as well as Jaccard of 0.77 and 0.72 for binary (Ophiuroidea detection) and multiclass segmentation tasks, respectively. Effective, accurate and fast segmentation of the seabed mosaic can be achieved if sufficient training annotations are available for the DL model. Further development is envisaged to enable tweaking capabilities for segmentation results as the post-processing phase. |
format |
Conference Object |
author |
Buškus, Kazimieras Vaičiukynas, Evaldas Verikas, Antanas Medelytė, Saulė Olenin, Sergej |
author_facet |
Buškus, Kazimieras Vaičiukynas, Evaldas Verikas, Antanas Medelytė, Saulė Olenin, Sergej |
author_sort |
Buškus, Kazimieras |
title |
Prototype framework for deep learning driven semantic segmentation of Arctic seabed imagery |
title_short |
Prototype framework for deep learning driven semantic segmentation of Arctic seabed imagery |
title_full |
Prototype framework for deep learning driven semantic segmentation of Arctic seabed imagery |
title_fullStr |
Prototype framework for deep learning driven semantic segmentation of Arctic seabed imagery |
title_full_unstemmed |
Prototype framework for deep learning driven semantic segmentation of Arctic seabed imagery |
title_sort |
prototype framework for deep learning driven semantic segmentation of arctic seabed imagery |
publishDate |
2021 |
url |
https://vb.ktu.edu/KTU:ELABAPDB89723914&prefLang=en_US |
geographic |
Arctic |
geographic_facet |
Arctic |
genre |
Arctic Arctic Spitsbergen |
genre_facet |
Arctic Arctic Spitsbergen |
op_source |
Arctic science: Arctic Change 2020 conference book of abstracts = Compilation de Résumés pour la Conférence Arctic Change 2020: virtual event hosted by Université Laval, Québec, Canada, Ottawa, Ontario : Canadian science publishing, 2021, vol. 7, iss. 1, ID: 463, p. 85 ISSN 2368-7460 |
op_relation |
info:eu-repo/semantics/altIdentifier/doi/10.1139/as-2021-0001 https://epubl.ktu.edu/object/elaba:89723914/89723914.pdf https://vb.ktu.edu/KTU:ELABAPDB89723914&prefLang=en_US |
op_rights |
info:eu-repo/semantics/openAccess |
op_doi |
https://doi.org/10.1139/as-2021-0001 |
container_title |
Arctic Science |
container_volume |
7 |
container_issue |
1 |
container_start_page |
3 |
op_container_end_page |
135 |
_version_ |
1766294541496745984 |