Afficher la notice abrégée

Image based species identification of Globodera quarantine nematodes using computer vision and deep learning

hal.structure.identifierInstitut de Génétique, Environnement et Protection des Plantes [IGEPP]
hal.structure.identifierUnité de Nématologie [LSV Rennes]
dc.contributor.authorTHEVENOUX, Romain
hal.structure.identifierModélisation Mathématique pour l'Oncologie [MONC]
dc.contributor.authorLE, Van-Linh
hal.structure.identifierInstitut de Génétique, Environnement et Protection des Plantes [IGEPP]
hal.structure.identifierUnité de Nématologie [LSV Rennes]
dc.contributor.authorVILLESSÈCHE, Heloïse
hal.structure.identifierUnité de Nématologie [LSV Rennes]
dc.contributor.authorBUISSON, Alain
hal.structure.identifierLaboratoire Bordelais de Recherche en Informatique [LaBRI]
dc.contributor.authorBEURTON-AIMAR, Marie
hal.structure.identifierInstitut de Génétique, Environnement et Protection des Plantes [IGEPP]
dc.contributor.authorGRENIER, Eric
hal.structure.identifierUnité de Nématologie [LSV Rennes]
dc.contributor.authorFOLCHER, Laurent
hal.structure.identifierInstitut de Génétique, Environnement et Protection des Plantes [IGEPP]
dc.contributor.authorPARISEY, Nicolas
dc.date.accessioned2024-04-04T02:33:55Z
dc.date.available2024-04-04T02:33:55Z
dc.date.issued2021-07
dc.identifier.issn0168-1699
dc.identifier.urihttps://oskar-bordeaux.fr/handle/20.500.12278/190518
dc.description.abstractEnIdentification of plant parasitic nematode species is usually achieved following morphobiometric analysis, which requires a certain level of expertise and remains time consuming. Moreover, molecular and morphological discrimination of a number of emergent or cryptic species is sometimes difficult. Finding a way to achieve morphological characterisation quickly and accurately would greatly advance nematology science. Here, we developed a complete method in order to identify the two quarantine nematode species Globodera pallida and Globodera rostochiensis. First, we chose discriminative metrics on the stylet of nematodes that are able to be used by algorithms in order to build an automated process. Second, we used a custom computer vision algorithm (CCVA) and a convolutional neural network (CNN) to measure our metrics of interest. Third, we compared the CCVA and CNN predictions and their discriminative power to distinguish closely related species. Results show accurate identification of G. pallida and G. rostochiensis with the two methods, despite small-scale divergence (one to five µm depending on the metric used). However, the error rate is higher for Globodera mexicana, suggesting that the algorithms are too specific. Nonetheless, these methods represent a promising novel approach to automated morphological identification of nematodes and Globodera species in particular.
dc.language.isoen
dc.publisherElsevier
dc.rights.urihttp://creativecommons.org/licenses/by-nc/
dc.subject.enAutomation
dc.subject.enLandmarks
dc.subject.enMachine learning
dc.subject.enMorphometrics
dc.subject.enPotato cyst nematode
dc.subject.enNematode taxonomy
dc.title.enImage based species identification of Globodera quarantine nematodes using computer vision and deep learning
dc.typeArticle de revue
dc.identifier.doi10.1016/j.compag.2021.106058
dc.subject.halSciences du Vivant [q-bio]
bordeaux.journalComputers and Electronics in Agriculture
bordeaux.volume186
bordeaux.hal.laboratoriesInstitut de Mathématiques de Bordeaux (IMB) - UMR 5251*
bordeaux.institutionUniversité de Bordeaux
bordeaux.institutionBordeaux INP
bordeaux.institutionCNRS
bordeaux.peerReviewedoui
hal.identifierhal-03319310
hal.version1
hal.popularnon
hal.audienceInternationale
hal.origin.linkhttps://hal.archives-ouvertes.fr//hal-03319310v1
bordeaux.COinSctx_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.jtitle=Computers%20and%20Electronics%20in%20Agriculture&rft.date=2021-07&rft.volume=186&rft.eissn=0168-1699&rft.issn=0168-1699&rft.au=THEVENOUX,%20Romain&LE,%20Van-Linh&VILLESS%C3%88CHE,%20Helo%C3%AFse&BUISSON,%20Alain&BEURTON-AIMAR,%20Marie&rft.genre=article


Fichier(s) constituant ce document

FichiersTailleFormatVue

Il n'y a pas de fichiers associés à ce document.

Ce document figure dans la(les) collection(s) suivante(s)

Afficher la notice abrégée