FN Archimer Export Format
PT J
TI Complex data labeling with deep learning methods: Lessons from fisheries acoustics
BT 
AF Sarr, J.M.A.
   Brochier, T.
   Brehmer, Patrice
   Perrot, Yannick
   Bah, A.
   Sarré, A.
   Jeyid, M.A.
   Sidibeh, M.
   El Ayoubi, S.
AS 1:1,2;2:1,2;3:3,4;4:3;5:1,2;6:4;7:5;8:6;9:7;
FF 1:;2:;3:;4:;5:;6:;7:;8:;9:;
C1 Université Cheikh Anta Diop de Dakar UCAD, Ecole Supérieure Polytechnique, BP 15915, Dakar Fann, Senegal
   IRD, Sorbonne Université, UMMISCO, F-93143, Bondy, France
   IRD, Univ Brest, CNRS, Ifremer, LEMAR, Plouzané, France
   ISRA/CRODT, Pole de recherche de Hann, BP2241, Dakar, Senegal
   IMROP, BP22, Nouadhibou, Mauritania
   Fisheries Department (FD), Marina Bay, Banjul, The Gambia
   INRH, Anza 80000, Agadir, Morocco
C2 UNIV CHEIKH ANTA DIOP (UCAD), SENEGAL
   IRD, FRANCE
   IRD, FRANCE
   CRODT, SENEGAL
   IMROP, MAURITANIA
   FISHERIES DPT, GAMBIA
   INRH, MOROCCO
UM LEMAR
IN WOS Cotutelle UMR
   copubli-france
   copubli-int-hors-europe
   copubli-sud
IF 5.911
TC 10
UR https://archimer.ifremer.fr/doc/00653/76545/77658.pdf
LA English
DT Article
DE ;Fisheries acoustics;Machine learning;Neural network;Active acoustics;Labeling process;Bottom correction
AB Quantitative and qualitative analysis of acoustic backscattered signals from the seabed bottom to the sea surface is used worldwide for fish stocks assessment and marine ecosystem monitoring. Huge amounts of raw data are collected yet require tedious expert labeling. This paper focuses on a case study where the ground truth labels are non-obvious: echograms labeling, which is time-consuming and critical for the quality of fisheries and ecological analysis. We investigate how these tasks can benefit from supervised learning algorithms and demonstrate that convolutional neural networks trained with non-stationary datasets can be used to stress parts of a new dataset needing human expert correction. Further development of this approach paves the way toward a standardization of the labeling process in fisheries acoustics and is a good case study for non-obvious data labeling processes. 
PY 2021
PD MAR
SO Isa Transactions
SN 0019-0578
PU Elsevier BV
VL 109
UT 000618971000011
BP 113
EP 125
DI 10.1016/j.isatra.2020.09.018
ID 76545
ER

EF