Workflow and convolutional neural network for automated identification of animal sounds

The use of passive acoustic monitoring in wildlife ecology has increased dramatically in recent years as researchers take advantage of improvements in autonomous recording units and analytical methods. These technologies have allowed researchers to collect large quantities of acoustic data which mus...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Zachary J. Ruff, Damon B. Lesmeister, Cara L. Appel, Christopher M. Sullivan
Formato:	article
Lenguaje:	EN
Publicado:	Elsevier 2021
Materias:	Bioacoustics Machine learning Wildlife Ecology Passive acoustic monitoring Artificial intelligence QH540-549.5
Acceso en línea:	https://doaj.org/article/1ba426ca651b4d3e9f4c312db2aa0988
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:1ba426ca651b4d3e9f4c312db2aa0988
record_format	dspace
spelling	oai:doaj.org-article:1ba426ca651b4d3e9f4c312db2aa09882021-12-01T04:45:53ZWorkflow and convolutional neural network for automated identification of animal sounds1470-160X10.1016/j.ecolind.2021.107419https://doaj.org/article/1ba426ca651b4d3e9f4c312db2aa09882021-05-01T00:00:00Zhttp://www.sciencedirect.com/science/article/pii/S1470160X21000844https://doaj.org/toc/1470-160XThe use of passive acoustic monitoring in wildlife ecology has increased dramatically in recent years as researchers take advantage of improvements in autonomous recording units and analytical methods. These technologies have allowed researchers to collect large quantities of acoustic data which must then be processed to extract meaningful information, e.g. target species detections. A persistent issue in acoustic monitoring is the challenge of efficiently automating the detection of species of interest, and deep learning has emerged as a powerful approach to accomplish this task. Here we report on the development and application of a deep convolutional neural network for the automated detection of 14 forest-adapted birds and mammals by classifying spectrogram images generated from short audio clips. The neural network performed well for most species, with precision exceeding 90% and recall exceeding 50% at high score thresholds, indicating high power to detect these species when they were present and vocally active, combined with a low proportion of false positives. We describe a multi-step workflow that integrates this neural network to efficiently process large volumes of audio data with a combination of automated detection and human review. This workflow reduces the necessary human effort by > 99% compared to full manual review of the data. As an optional component of this workflow, we developed a graphical interface for the neural network that can be run through RStudio using the Shiny package, creating a portable and user-friendly way for field biologists and managers to efficiently process audio data and detect these target species close to the point of collection and with minimal delays using consumer-grade computers.Zachary J. RuffDamon B. LesmeisterCara L. AppelChristopher M. SullivanElsevierarticleBioacousticsMachine learningWildlifeEcologyPassive acoustic monitoringArtificial intelligenceEcologyQH540-549.5ENEcological Indicators, Vol 124, Iss , Pp 107419- (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Bioacoustics Machine learning Wildlife Ecology Passive acoustic monitoring Artificial intelligence Ecology QH540-549.5
spellingShingle	Bioacoustics Machine learning Wildlife Ecology Passive acoustic monitoring Artificial intelligence Ecology QH540-549.5 Zachary J. Ruff Damon B. Lesmeister Cara L. Appel Christopher M. Sullivan Workflow and convolutional neural network for automated identification of animal sounds
description	The use of passive acoustic monitoring in wildlife ecology has increased dramatically in recent years as researchers take advantage of improvements in autonomous recording units and analytical methods. These technologies have allowed researchers to collect large quantities of acoustic data which must then be processed to extract meaningful information, e.g. target species detections. A persistent issue in acoustic monitoring is the challenge of efficiently automating the detection of species of interest, and deep learning has emerged as a powerful approach to accomplish this task. Here we report on the development and application of a deep convolutional neural network for the automated detection of 14 forest-adapted birds and mammals by classifying spectrogram images generated from short audio clips. The neural network performed well for most species, with precision exceeding 90% and recall exceeding 50% at high score thresholds, indicating high power to detect these species when they were present and vocally active, combined with a low proportion of false positives. We describe a multi-step workflow that integrates this neural network to efficiently process large volumes of audio data with a combination of automated detection and human review. This workflow reduces the necessary human effort by > 99% compared to full manual review of the data. As an optional component of this workflow, we developed a graphical interface for the neural network that can be run through RStudio using the Shiny package, creating a portable and user-friendly way for field biologists and managers to efficiently process audio data and detect these target species close to the point of collection and with minimal delays using consumer-grade computers.
format	article
author	Zachary J. Ruff Damon B. Lesmeister Cara L. Appel Christopher M. Sullivan
author_facet	Zachary J. Ruff Damon B. Lesmeister Cara L. Appel Christopher M. Sullivan
author_sort	Zachary J. Ruff
title	Workflow and convolutional neural network for automated identification of animal sounds
title_short	Workflow and convolutional neural network for automated identification of animal sounds
title_full	Workflow and convolutional neural network for automated identification of animal sounds
title_fullStr	Workflow and convolutional neural network for automated identification of animal sounds
title_full_unstemmed	Workflow and convolutional neural network for automated identification of animal sounds
title_sort	workflow and convolutional neural network for automated identification of animal sounds
publisher	Elsevier
publishDate	2021
url	https://doaj.org/article/1ba426ca651b4d3e9f4c312db2aa0988
work_keys_str_mv	AT zacharyjruff workflowandconvolutionalneuralnetworkforautomatedidentificationofanimalsounds AT damonblesmeister workflowandconvolutionalneuralnetworkforautomatedidentificationofanimalsounds AT caralappel workflowandconvolutionalneuralnetworkforautomatedidentificationofanimalsounds AT christophermsullivan workflowandconvolutionalneuralnetworkforautomatedidentificationofanimalsounds
_version_	1718405766722879488

Workflow and convolutional neural network for automated identification of animal sounds

Ejemplares similares