A Deep Learning Model to Recognize and Quantitatively Analyze Cold Seep Substrates and the Dominant Associated Species

Characterizing habitats and species distribution is important to understand the structure and function of cold seep ecosystems. This paper develops a deep learning model for the fast and accurate recognition and classification of substrates and the dominant associated species in cold seeps. Consider...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Haining Wang, Xiaoxue Fu, Chengqian Zhao, Zhendong Luan, Chaolun Li
Formato: article
Lenguaje:EN
Publicado: Frontiers Media S.A. 2021
Materias:
FPN
Q
Acceso en línea:https://doaj.org/article/6fba8cbba4324a17888f93f7fa5541c4
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:6fba8cbba4324a17888f93f7fa5541c4
record_format dspace
spelling oai:doaj.org-article:6fba8cbba4324a17888f93f7fa5541c42021-12-01T02:02:30ZA Deep Learning Model to Recognize and Quantitatively Analyze Cold Seep Substrates and the Dominant Associated Species2296-774510.3389/fmars.2021.775433https://doaj.org/article/6fba8cbba4324a17888f93f7fa5541c42021-11-01T00:00:00Zhttps://www.frontiersin.org/articles/10.3389/fmars.2021.775433/fullhttps://doaj.org/toc/2296-7745Characterizing habitats and species distribution is important to understand the structure and function of cold seep ecosystems. This paper develops a deep learning model for the fast and accurate recognition and classification of substrates and the dominant associated species in cold seeps. Considering the dense distribution of the dominant associated species and small objects caused by overlap in cold seeps, the feature pyramid network (FPN) embed into the faster region-convolutional neural network (R-CNN) was used to detect large-scale changes and small missing objects without increasing the number of calculations. We applied three classifiers (Faster R-CNN + FPN for mussel beds, lobster clusters and biological mixing, CNN for shell debris and exposed authigenic carbonates, and VGG16 for reduced sediments and muddy bottom) to improve the recognition accuracy of substrates. The model’s results were manually verified using images obtained in the Formosa cold seep during a 2016 cruise. The recognition accuracy of the two dominant species, e.g., Gigantidas platifrons and Munidopsidae could be 70.85 and 56.16%, respectively. Seven subcategories of substrates were also classified with a mean accuracy of 74.87%. The developed model is a promising tool for the fast and accurate characterization of substrates and epifauna in cold seeps, which is crucial for large-scale quantitative analyses.Haining WangHaining WangHaining WangXiaoxue FuChengqian ZhaoZhendong LuanZhendong LuanZhendong LuanZhendong LuanChaolun LiChaolun LiChaolun LiChaolun LiFrontiers Media S.A.articlecold seepsubstratesepifaunaFaster R-CNNFPNVGG16ScienceQGeneral. Including nature conservation, geographical distributionQH1-199.5ENFrontiers in Marine Science, Vol 8 (2021)
institution DOAJ
collection DOAJ
language EN
topic cold seep
substrates
epifauna
Faster R-CNN
FPN
VGG16
Science
Q
General. Including nature conservation, geographical distribution
QH1-199.5
spellingShingle cold seep
substrates
epifauna
Faster R-CNN
FPN
VGG16
Science
Q
General. Including nature conservation, geographical distribution
QH1-199.5
Haining Wang
Haining Wang
Haining Wang
Xiaoxue Fu
Chengqian Zhao
Zhendong Luan
Zhendong Luan
Zhendong Luan
Zhendong Luan
Chaolun Li
Chaolun Li
Chaolun Li
Chaolun Li
A Deep Learning Model to Recognize and Quantitatively Analyze Cold Seep Substrates and the Dominant Associated Species
description Characterizing habitats and species distribution is important to understand the structure and function of cold seep ecosystems. This paper develops a deep learning model for the fast and accurate recognition and classification of substrates and the dominant associated species in cold seeps. Considering the dense distribution of the dominant associated species and small objects caused by overlap in cold seeps, the feature pyramid network (FPN) embed into the faster region-convolutional neural network (R-CNN) was used to detect large-scale changes and small missing objects without increasing the number of calculations. We applied three classifiers (Faster R-CNN + FPN for mussel beds, lobster clusters and biological mixing, CNN for shell debris and exposed authigenic carbonates, and VGG16 for reduced sediments and muddy bottom) to improve the recognition accuracy of substrates. The model’s results were manually verified using images obtained in the Formosa cold seep during a 2016 cruise. The recognition accuracy of the two dominant species, e.g., Gigantidas platifrons and Munidopsidae could be 70.85 and 56.16%, respectively. Seven subcategories of substrates were also classified with a mean accuracy of 74.87%. The developed model is a promising tool for the fast and accurate characterization of substrates and epifauna in cold seeps, which is crucial for large-scale quantitative analyses.
format article
author Haining Wang
Haining Wang
Haining Wang
Xiaoxue Fu
Chengqian Zhao
Zhendong Luan
Zhendong Luan
Zhendong Luan
Zhendong Luan
Chaolun Li
Chaolun Li
Chaolun Li
Chaolun Li
author_facet Haining Wang
Haining Wang
Haining Wang
Xiaoxue Fu
Chengqian Zhao
Zhendong Luan
Zhendong Luan
Zhendong Luan
Zhendong Luan
Chaolun Li
Chaolun Li
Chaolun Li
Chaolun Li
author_sort Haining Wang
title A Deep Learning Model to Recognize and Quantitatively Analyze Cold Seep Substrates and the Dominant Associated Species
title_short A Deep Learning Model to Recognize and Quantitatively Analyze Cold Seep Substrates and the Dominant Associated Species
title_full A Deep Learning Model to Recognize and Quantitatively Analyze Cold Seep Substrates and the Dominant Associated Species
title_fullStr A Deep Learning Model to Recognize and Quantitatively Analyze Cold Seep Substrates and the Dominant Associated Species
title_full_unstemmed A Deep Learning Model to Recognize and Quantitatively Analyze Cold Seep Substrates and the Dominant Associated Species
title_sort deep learning model to recognize and quantitatively analyze cold seep substrates and the dominant associated species
publisher Frontiers Media S.A.
publishDate 2021
url https://doaj.org/article/6fba8cbba4324a17888f93f7fa5541c4
work_keys_str_mv AT hainingwang adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT hainingwang adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT hainingwang adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT xiaoxuefu adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chengqianzhao adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT zhendongluan adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT zhendongluan adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT zhendongluan adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT zhendongluan adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chaolunli adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chaolunli adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chaolunli adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chaolunli adeeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT hainingwang deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT hainingwang deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT hainingwang deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT xiaoxuefu deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chengqianzhao deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT zhendongluan deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT zhendongluan deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT zhendongluan deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT zhendongluan deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chaolunli deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chaolunli deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chaolunli deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
AT chaolunli deeplearningmodeltorecognizeandquantitativelyanalyzecoldseepsubstratesandthedominantassociatedspecies
_version_ 1718405964936249344