FN Archimer Export Format
PT J
TI Convolutional neural networks for hydrothermal vents substratum classification: An introspective study
BT 
AF SOTO VEGA, Pedro Juan
   Papadakis, Panagiotis
   Matabos, Marjolaine
   Van Audenhaege, Loic
   Ramiere, Annah
   Sarrazin, Jozee
   Pedro da Costa, Gilson Alexandre Ostwald
AS 1:1;2:2;3:3;4:4;5:3;6:3;7:5;
FF 1:;2:;3:PDG-REM-BEEP-LEP;4:PDG-REM-BEEP-LEP;5:;6:PDG-REM-BEEP-LEP;7:;
C1 University Brest, LaTIM, INSERM, UMR 1101, 29200 Brest, France
   IMT Atlantique, Lab-STICC, UMR 6285, Team RAMBO, F-29238 Brest, France
   University Brest, CNRS, Ifremer, UMR6197 Biologie et Ecologie des Ecosystèmes marins Profonds, 29280 Plouzané, France.
   National Oceanography Center, Southampton, United Kingdom.
   Institute of Mathematics and Statistics, State University of Rio de Janeiro (UERJ), Rio de Janeiro, Brazil
C2 UBO, FRANCE
   IMT ATLANTIQUE, FRANCE
   IFREMER, FRANCE
   NOC, UK
   UNIV RIO DE JANEIRO, BRAZIL
SI BREST
SE PDG-REM-BEEP-LEP
UM BEEP-LM2E
IN WOS Ifremer UMR
   DOAJ
   copubli-france
   copubli-europe
   copubli-univ-france
   copubli-int-hors-europe
   copubli-sud
IF 5.1
TC 0
UR https://archimer.ifremer.fr/doc/00879/99050/108838.pdf
   https://archimer.ifremer.fr/doc/00879/99050/108839.jpg
LA English
DT Article
CR MOMARSAT : MONITORING THE MID ATLANTIC RIDGE
   MOMARSAT2018
BO L'Atalante
DE ;Image classification;Deep learning;Hydrothermal vents;Uncertainty analysis
AB The increasing availability of seabed images has created new opportunities and challenges for monitoring and better understanding the spatial distribution of fauna and substrata. To date, however, deep-sea substratum classification relies mostly on visual interpretation, which is costly, time-consuming, and prone to human bias or error. Motivated by the success of convolutional neural networks in learning semantically rich representations directly from images, this work investigates the application of state-of-the-art network architectures, originally employed in the classification of non-seabed images, for the task of hydrothermal vent substrata image classification. In assessing their potential, we conduct a study on the generalization, complementarity and human interpretability aspects of those architectures. Specifically, we independently trained deep learning models with the selected architectures using images obtained from three distinct sites within the Lucky-Strike vent field and assessed the models' performances on-site as well as off-site. To investigate complementarity, we evaluated a classification decision committee (CDC) built as an ensemble of networks in which individual predictions were fused through a majority voting scheme. The experimental results demonstrated the suitability of the deep learning models for deep-sea substratum classification, attaining accuracies reaching up to 80% in terms of F1-score. Finally, by further investigating the classification uncertainty computed from the set of individual predictions of the CDC, we describe a semiautomatic framework for human annotation, which prescribes visual inspection of only the images with high uncertainty. Overall, the results demonstrated that high accuracy values of over 90% F1-score can be obtained with the framework, with a small amount of human intervention. 
PY 2024
PD MAY
SO Ecological Informatics
SN 1574-9541
PU Elsevier BV
VL 80
UT 001195424100001
DI 10.1016/j.ecoinf.2024.102535
ID 99050
ER

EF