Frog calling activity detection using lightweight CNN with multi-view spectrogram: A case study on Kroombit tinker frog

Frogs play an important role in ecological systems, while frog species across the globe are threatened and declining. Therefore, it is valuable to estimate the frog population based on an intelligent computer system. Due to the success of deep learning (DL) in various pattern recognition tasks, prev...

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Jie Xie, Mingying Zhu, Kai Hu, Jinglan Zhang, Harry Hines, Ya Guo
Formato: article
Lenguaje:EN
Publicado: Elsevier 2022
Materias:
Acceso en línea:https://doaj.org/article/67e7710c4b124390be626699700b55f3
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
id oai:doaj.org-article:67e7710c4b124390be626699700b55f3
record_format dspace
spelling oai:doaj.org-article:67e7710c4b124390be626699700b55f32021-11-20T05:15:23ZFrog calling activity detection using lightweight CNN with multi-view spectrogram: A case study on Kroombit tinker frog2666-827010.1016/j.mlwa.2021.100202https://doaj.org/article/67e7710c4b124390be626699700b55f32022-03-01T00:00:00Zhttp://www.sciencedirect.com/science/article/pii/S2666827021001018https://doaj.org/toc/2666-8270Frogs play an important role in ecological systems, while frog species across the globe are threatened and declining. Therefore, it is valuable to estimate the frog population based on an intelligent computer system. Due to the success of deep learning (DL) in various pattern recognition tasks, previous studies have used DL-based methods for frog call analysis. However, the performance of DL-based systems is highly affected by their input (feature representation). In this study, we develop a frog calling activity detection system for continuous field recordings using a light convolutional neural network (CNN) with multi-view spectrograms. To be specific, a sliding window is first applied to continuous recordings for obtaining audio segments with a fixed duration. Then, the background noise is filtered out. Next, a multi-view spectrogram is used for characterizing those segments, which has more distinctive information than a single-view spectrogram. Finally, a lightweight CNN model is used for the detection of frog calling activity with a twin loss, where different train and test sets are used to validate the model’s robustness. Our experimental results indicate that the highest macro F1-score was 99.6 ± 0.2 and 96.4 ± 2.0 using 2016 and 2017 as the train data respectively, where CNN-GAP is used as the model with multi-view spectrogram as the input.Jie XieMingying ZhuKai HuJinglan ZhangHarry HinesYa GuoElsevierarticleBioacoustic signal activity detectionMulti-view spectrogramLightweight CNNLoss functionCyberneticsQ300-390Electronic computers. Computer scienceQA75.5-76.95ENMachine Learning with Applications, Vol 7, Iss , Pp 100202- (2022)
institution DOAJ
collection DOAJ
language EN
topic Bioacoustic signal activity detection
Multi-view spectrogram
Lightweight CNN
Loss function
Cybernetics
Q300-390
Electronic computers. Computer science
QA75.5-76.95
spellingShingle Bioacoustic signal activity detection
Multi-view spectrogram
Lightweight CNN
Loss function
Cybernetics
Q300-390
Electronic computers. Computer science
QA75.5-76.95
Jie Xie
Mingying Zhu
Kai Hu
Jinglan Zhang
Harry Hines
Ya Guo
Frog calling activity detection using lightweight CNN with multi-view spectrogram: A case study on Kroombit tinker frog
description Frogs play an important role in ecological systems, while frog species across the globe are threatened and declining. Therefore, it is valuable to estimate the frog population based on an intelligent computer system. Due to the success of deep learning (DL) in various pattern recognition tasks, previous studies have used DL-based methods for frog call analysis. However, the performance of DL-based systems is highly affected by their input (feature representation). In this study, we develop a frog calling activity detection system for continuous field recordings using a light convolutional neural network (CNN) with multi-view spectrograms. To be specific, a sliding window is first applied to continuous recordings for obtaining audio segments with a fixed duration. Then, the background noise is filtered out. Next, a multi-view spectrogram is used for characterizing those segments, which has more distinctive information than a single-view spectrogram. Finally, a lightweight CNN model is used for the detection of frog calling activity with a twin loss, where different train and test sets are used to validate the model’s robustness. Our experimental results indicate that the highest macro F1-score was 99.6 ± 0.2 and 96.4 ± 2.0 using 2016 and 2017 as the train data respectively, where CNN-GAP is used as the model with multi-view spectrogram as the input.
format article
author Jie Xie
Mingying Zhu
Kai Hu
Jinglan Zhang
Harry Hines
Ya Guo
author_facet Jie Xie
Mingying Zhu
Kai Hu
Jinglan Zhang
Harry Hines
Ya Guo
author_sort Jie Xie
title Frog calling activity detection using lightweight CNN with multi-view spectrogram: A case study on Kroombit tinker frog
title_short Frog calling activity detection using lightweight CNN with multi-view spectrogram: A case study on Kroombit tinker frog
title_full Frog calling activity detection using lightweight CNN with multi-view spectrogram: A case study on Kroombit tinker frog
title_fullStr Frog calling activity detection using lightweight CNN with multi-view spectrogram: A case study on Kroombit tinker frog
title_full_unstemmed Frog calling activity detection using lightweight CNN with multi-view spectrogram: A case study on Kroombit tinker frog
title_sort frog calling activity detection using lightweight cnn with multi-view spectrogram: a case study on kroombit tinker frog
publisher Elsevier
publishDate 2022
url https://doaj.org/article/67e7710c4b124390be626699700b55f3
work_keys_str_mv AT jiexie frogcallingactivitydetectionusinglightweightcnnwithmultiviewspectrogramacasestudyonkroombittinkerfrog
AT mingyingzhu frogcallingactivitydetectionusinglightweightcnnwithmultiviewspectrogramacasestudyonkroombittinkerfrog
AT kaihu frogcallingactivitydetectionusinglightweightcnnwithmultiviewspectrogramacasestudyonkroombittinkerfrog
AT jinglanzhang frogcallingactivitydetectionusinglightweightcnnwithmultiviewspectrogramacasestudyonkroombittinkerfrog
AT harryhines frogcallingactivitydetectionusinglightweightcnnwithmultiviewspectrogramacasestudyonkroombittinkerfrog
AT yaguo frogcallingactivitydetectionusinglightweightcnnwithmultiviewspectrogramacasestudyonkroombittinkerfrog
_version_ 1718419486130831360