RFIDeep: Unfolding the potential of deep learning for radio‐frequency identification

Article Article scientifique Article RFIDeep: Unfolding the potential of deep learning for radio‐frequency identification eng 2023-08-22 2023-08-22 author Bardon Gaël 0000-0002-0243-9178 Centre Scientifique de Monaco, Département de Biologie Polaire Monaco Principality of Monaco author Cristofari Robin 0000-0001-8430-7453 University of Turku, Department of Biology Turku Finland author Winterl Alexander 0000-0003-0688-9317 Friedrich‐Alexander‐University Erlangen‐Nürnberg, Department of Physics Erlangen Germany author Barracho Téo 0000-0002-8742-4062 Université de Strasbourg, CNRS, IPHC UMR 7178 Strasbourg France author Benoiste Marine Université de Strasbourg, CNRS, IPHC UMR 7178 Strasbourg France author Ceresa Claire Centre Scientifique de Monaco, Département de Biologie Polaire Monaco Principality of Monaco author Chatelain Nicolas 0000-0002-1346-3595 Université de Strasbourg, CNRS, IPHC UMR 7178 Strasbourg France author Courtecuisse Julien 0000-0002-0179-379X Université de Strasbourg, CNRS, IPHC UMR 7178 Strasbourg France author Fernandes Flávia A. N. 0000-0001-7897-6976 Université de Strasbourg, CNRS, IPHC UMR 7178 Strasbourg France author Gauthier-Clerc Michel University of Geneva, Faculty of Sciences Geneva Switzerland 986678 Administration de la faculté des sciences (SIAD) author Gendner Jean‐Paul Université de Strasbourg, CNRS, IPHC UMR 7178 Strasbourg France author Handrich Yves 0000-0002-4478-5562 Université de Strasbourg, CNRS, IPHC UMR 7178 Strasbourg France author Houstin Aymeric 0000-0001-8676-819X Centre Scientifique de Monaco, Département de Biologie Polaire Monaco Principality of Monaco author Krellenstein Adélie Université de Strasbourg, CNRS, IPHC UMR 7178 Strasbourg France author Lecomte Nicolas 0000-0002-8473-5375 University of Moncton, Canada Research Chair in Polar and Boreal Ecology and Centre d'Études Nordiques, Department of Biology Moncton New Brunswick Canada author Salmon Charles‐Edouard Beefutures Nantes France author Trucchi Emiliano 0000-0002-1270-5273 Marche Polytechnic University, Department of Life and Environmental Sciences Ancona Italy author Vallas Benoit Université de Strasbourg, CNRS, IPHC UMR 7178 Strasbourg France author Wong Emily M. 0000-0003-1445-1614 Woods Hole Oceanographic Institution, Applied Ocean Physics and Engineering Department Woods Hole Massachusetts USA author Zitterbart Daniel P. 0000-0001-9429-4350 Friedrich‐Alexander‐University Erlangen‐Nürnberg, Department of Physics Erlangen Germany author Le Bohec Céline 0000-0003-0149-6477 Centre Scientifique de Monaco, Département de Biologie Polaire Monaco Principality of Monaco 2041-210X.14187 10.1111/2041-210X.14187 2041-210X https://besjournals.onlinelibrary.wiley.com/doi/10.1111/2041-210X.14187 Automatic monitoring of wildlife is becoming a critical tool in the field of ecology. In particular, Radio‐Frequency IDentification (RFID) is now a widespread technology to assess the phenology, breeding and survival of many species. While RFID produces massive datasets, no established fast and accurate methods are yet available for this type of data processing. Deep learning approaches have been used to overcome similar problems in other scientific fields and hence might hold the potential to overcome these analytical challenges and unlock the full potential of RFID studies. We present a deep learning workflow, coined “RFIDeep”, to derive ecological features, such as breeding status and outcome, from RFID mark‐recapture data. To demonstrate the performance of RFIDeep with complex datasets, we used a long‐term automatic monitoring of a long‐lived seabird that breeds in densely packed colonies, hence with many daily entries and exits. To determine individual breeding status and phenology and for each breeding season, we first developed a one‐dimensional convolution neural network (1D‐CNN) architecture. Second, to account for variance in breeding phenology and technical limitations of field data acquisition, we built a new data augmentation step mimicking a shift in breeding dates and missing RFID detections, a common issue with RFIDs. Third, to identify the segments of the breeding activity used during classification, we also included a visualisation tool, which allows users to understand what is usually considered a “black box” step of deep learning. With these three steps, we achieved a high accuracy for all breeding parameters: breeding status accuracy = 96.3%; phenological accuracy = 86.9%; and breeding success accuracy = 97.3%. RFIDeep has unfolded the potential of artificial intelligence for tracking changes in animal populations, multiplying the benefit of automated mark‐recapture monitoring of undisturbed wildlife populations. RFIDeep is an open source code to facilitate the use, adaptation, or enhancement of RFID data in a wide variety of species. In addition to a tremendous time saving for analysing these large datasets, our study shows the capacities of CNN models to autonomously detect ecologically meaningful patterns in data through visualisation techniques, which are seldom used in ecology. Dewey 500 Academy of Finland Genomic determinants of Ageing and Cancer avoidance in Elephants 331320 FI 2 Faculté des sciences 1 false Methods in ecology and evolution e4c1aba9-970f-4b84-b63b-3d23b1a7e566 Article (Published version) Public Methods Ecol Evol - 2023 - Bardon.pdf 2797615 application/pdf CC BY-4.0 aa024f0c042d7951a0e5ea6695decb764dc1c6ada97ff0378b4016c761cdbf1e openAccess Faculté des sciences

Title	Published in	Access level	OA Policy	Year	Views	Downloads

RFIDeep: Unfolding the potential of deep learning for radio‐frequency identification	Methods in ecology and evolution			2023	30	11

Archive ouverte UNIGE

Gauthier-Clerc, Michel

account_balance Affiliation entities

Affiliation entities