@article{51d0bdfe6a754103bbe18c220144bdb8,
title = "ANIMAL-SPOT enables animal-independent signal detection and classification using deep learning",
abstract = "Bioacoustic research spans a wide range of biological questions and applications, relying on identification of target species or smaller acoustic units, such as distinct call types. However, manually identifying the signal of interest is time-intensive, error-prone, and becomes unfeasible with large data volumes. Therefore, machine-driven algorithms are increasingly applied to various bioacoustic signal identification challenges. Nevertheless, biologists still have major difficulties trying to transfer existing animal- and/or scenario-related machine learning approaches to their specific animal datasets and scientific questions. This study presents an animal-independent, open-source deep learning framework, along with a detailed user guide. Three signal identification tasks, commonly encountered in bioacoustics research, were investigated: (1) target signal vs. background noise detection, (2) species classification, and (3) call type categorization. ANIMAL-SPOT successfully segmented human-annotated target signals in data volumes representing 10 distinct animal species and 1 additional genus, resulting in a mean test accuracy of 97.9%, together with an average area under the ROC curve (AUC) of 95.9%, when predicting on unseen recordings. Moreover, an average segmentation accuracy and F1-score of 95.4% was achieved on the publicly available BirdVox-Full-Night data corpus. In addition, multi-class species and call type classification resulted in 96.6% and 92.7% accuracy on unseen test data, as well as 95.2% and 88.4% regarding previous animal-specific machine-based detection excerpts. Furthermore, an Unweighted Average Recall (UAR) of 89.3% outperformed the multi-species classification baseline system of the ComParE 2021 Primate Sub-Challenge. Besides animal independence, ANIMAL-SPOT does not rely on expert knowledge or special computing resources, thereby making deep-learning-based bioacoustic signal identification accessible to a broad audience.",
keywords = "Acoustics, Algorithms, Animals, Area Under Curve, Deep Learning, Humans, Machine Learning",
author = "Christian Bergler and Smeele, {Simeon Q.} and Tyndel, {Stephen A.} and Alexander Barnhill and Ortiz, {Sara T.} and Kalan, {Ammie K.} and Cheng, {Rachael Xi} and Signe Brinkl{\o}v and Osiecka, {Anna N.} and Jakob Tougaard and Freja Jakobsen and Magnus Wahlberg and Elmar N{\"o}th and Andreas Maier and Klump, {Barbara C.}",
note = "Funding Information: We are grateful for support with—data collection of cockatiel data: Lucy Aplin and Gustavo Alarc{\'o}n-Nieto; data collection and annotations of killer whale data (orcalab.org61 ,62): Helena Symonds, Paul Spong, and Steven Ness4(formely UVIC); logistical support: Lucy Aplin (cockatiels, Sulphur-crested cockatoos, monk parakeets); John Martin, Anastasia Dalziell, and Justin Welbergen (Sulphur-crested cockatoos); Tim Wilder, Susan Vos and Fort McCoy Natural Resource Branch and Range Control (Blue- and Golden-winged warblers); access to field sites: Wisconsin and Illinois Department of Natural Resources (Blue- and Golden-winged warblers): The Bikuben Foundation and Mols Bjerge National Park (Pygmy pipistrelles); co-hosting the Heidelberg Academy of Sciences workshop: Jens Koblitz and Nora Carlson; design of the ANIMAL-SPOT network architecture image (part of Fig. 2): Michael Weber. We thank Joeri Zwerts for granting access to the ComParE-PRS primates dataset. All authors complied with the legislation in the respective countries were fieldwork was conducted. Ethical approval (where necessary) was obtained as follows: Data from cockatiels was collected by Lucy Aplin and Gustavo Alarc{\'o}n-Nieto under ethical permission from the Regierungspr{\"a}sidium Freiburg. Az. 35-9185.81/G-18/009; data from Sulphur-crested cockatoos was collected by Barbara Klump and John Martin under ethical permission from the Ethics Council of the Max Planck Society (application no. 2018_12; permit given to Lucy Aplin); data from Blue- and Golden-winged warblers was collected by Stephen Tyndel under ethical permission from the University of Illinois at Urbana-Champaign{\textquoteright}s IACUC committee (protocol #16022); data from chimpanzees was collected by Ammie Kalan under ethical permission from the Minist{\`e}re de la Recherche Scientifique, the Minist{\`e}re de l{\textquoteright}Environnement et des Eaux et For{\^e}ts, and the Office Ivorien des Parcs et Reserves in C{\^o}te d{\textquoteright}Ivoire (Ref: 11/MINEF/OIPR/DT/CAT). Research was conducted with funding from: German Research Council (DFG; grant MA-4898/18-1 to CB); Paul G. Allen Frontier{\textquoteright}s Group (CB); Max Planck Society (AKK; Mary Brooke McElreath funded SQS); Max Planck Society Independent Group Leader Fellowship to Lucy Aplin (funded SQS, SAT, BCK); International Max Planck Research School (IMPRS) for Organismal Biology (SQS, SAT, STO); German Academic Exchange Service (DAAD PhD scholarship to SAT); United States Department of Defense, Environmental Security Technology Certification Program (ESTCP grant #RC 201615 to Jinelle Sperry, Michael Ward and Brett Degregorio, U.S. Army Corps of Engineers, funded SAT); Animal Minds Project e.V. (STO); Carlsberg Foundation Semper Ardens grant to Peter Teglberg Madsen (funded SB); Danish Environmental Protection Agency (ANO, JT); Dansk Akustisk Selskab (FJ); University of Southern Denmark (Research grant to FJ); Office of Naval Research (MW); SDU Lighthouse Project (MW); Heidelberg Academy of Sciences (workshop grant to BCK). Funding Information: We are grateful for support with—data collection of cockatiel data: Lucy Aplin and Gustavo Alarc{\'o}n-Nieto; data collection and annotations of killer whale data (orcalab.org): Helena Symonds, Paul Spong, and Steven Ness (formely UVIC); logistical support: Lucy Aplin (cockatiels, Sulphur-crested cockatoos, monk parakeets); John Martin, Anastasia Dalziell, and Justin Welbergen (Sulphur-crested cockatoos); Tim Wilder, Susan Vos and Fort McCoy Natural Resource Branch and Range Control (Blue- and Golden-winged warblers); access to field sites: Wisconsin and Illinois Department of Natural Resources (Blue- and Golden-winged warblers): The Bikuben Foundation and Mols Bjerge National Park (Pygmy pipistrelles); co-hosting the Heidelberg Academy of Sciences workshop: Jens Koblitz and Nora Carlson; design of the ANIMAL-SPOT network architecture image (part of Fig. ): Michael Weber. We thank Joeri Zwerts for granting access to the ComParE-PRS primates dataset. All authors complied with the legislation in the respective countries were fieldwork was conducted. Ethical approval (where necessary) was obtained as follows: Data from cockatiels was collected by Lucy Aplin and Gustavo Alarc{\'o}n-Nieto under ethical permission from the Regierungspr{\"a}sidium Freiburg. Az. 35-9185.81/G-18/009; data from Sulphur-crested cockatoos was collected by Barbara Klump and John Martin under ethical permission from the Ethics Council of the Max Planck Society (application no. 2018_12; permit given to Lucy Aplin); data from Blue- and Golden-winged warblers was collected by Stephen Tyndel under ethical permission from the University of Illinois at Urbana-Champaign{\textquoteright}s IACUC committee (protocol #16022); data from chimpanzees was collected by Ammie Kalan under ethical permission from the Minist{\`e}re de la Recherche Scientifique, the Minist{\`e}re de l{\textquoteright}Environnement et des Eaux et For{\^e}ts, and the Office Ivorien des Parcs et Reserves in C{\^o}te d{\textquoteright}Ivoire (Ref: 11/MINEF/OIPR/DT/CAT). Research was conducted with funding from: German Research Council (DFG; grant MA-4898/18-1 to CB); Paul G. Allen Frontier{\textquoteright}s Group (CB); Max Planck Society (AKK; Mary Brooke McElreath funded SQS); Max Planck Society Independent Group Leader Fellowship to Lucy Aplin (funded SQS, SAT, BCK); International Max Planck Research School (IMPRS) for Organismal Biology (SQS, SAT, STO); German Academic Exchange Service (DAAD PhD scholarship to SAT); United States Department of Defense, Environmental Security Technology Certification Program (ESTCP grant #RC 201615 to Jinelle Sperry, Michael Ward and Brett Degregorio, U.S. Army Corps of Engineers, funded SAT); Animal Minds Project e.V. (STO); Carlsberg Foundation Semper Ardens grant to Peter Teglberg Madsen (funded SB); Danish Environmental Protection Agency (ANO, JT); Dansk Akustisk Selskab (FJ); University of Southern Denmark (Research grant to FJ); Office of Naval Research (MW); SDU Lighthouse Project (MW); Heidelberg Academy of Sciences (workshop grant to BCK). , Publisher Copyright: {\textcopyright} 2022, The Author(s).",
year = "2022",
month = dec,
doi = "10.1038/s41598-022-26429-y",
language = "English",
volume = "12",
pages = "21966",
journal = "Scientific Reports",
issn = "2045-2322",
publisher = "Nature Publishing Group",
}