Meta-Embedded Clustering (MEC): A new method for improving clustering quality in unlabeled bird sound datasets DOI Creative Commons
Joachim POUTARAUD, Jérôme Sueur,

Christophe Thébaud

et al.

Ecological Informatics, Journal Year: 2024, Volume and Issue: 82, P. 102687 - 102687

Published: June 15, 2024

In recent years, ecoacoustics has offered an alternative to traditional biodiversity monitoring techniques with the development of passive acoustic (PAM) systems allowing, among others, detect and identify species that are difficult by human observers, automatically. PAM typically generate large audio datasets, but using these infer ecologically meaningful information remains challenging. most cases, several thousand hours recordings need be manually labeled experts limiting operability systems. Based on developments meta-learning algorithms unsupervised learning techniques, we propose here Meta-Embedded Clustering (MEC), a new method high potential for improving clustering quality in unlabeled bird sound datasets. MEC is organized two main steps, with: (a) fine-tuning pretrained convolutional neural network (CNN) backbone different pseudo-labeled data, (b) manually-labeled sounds latent space based vector embeddings extracted from fine-tuned CNN. The significantly enhanced average performance less than 1% more 80%, greatly outperforming approach relying solely CNN features general neotropical database. However, this came cost excluding portion data categorized as noise. By should facilitate work ecoacousticians managing units song/call clustered according their similarities, identifying clusters undetected approaches.

Language: Английский

Soundscape analysis reveals fine ecological differences among coral reef habitats DOI Creative Commons

Juan Carlos Azofeifa-Solano,

Miles Parsons, Rohan M. Brooker

et al.

Ecological Indicators, Journal Year: 2025, Volume and Issue: 171, P. 113120 - 113120

Published: Jan. 23, 2025

Language: Английский

Citations

1
Temporal patterns in Malaysian rainforest soundscapes demonstrated using acoustic indices and deep embeddings trained on time-of-day estimation DOI Open Access
Yen Yi Loo,

Mei Yi Lee,

Samien Shaheed

et al.

The Journal of the Acoustical Society of America, Journal Year: 2025, Volume and Issue: 157(1), P. 1 - 16

Published: Jan. 1, 2025

Rapid urban development impacts the integrity of tropical ecosystems on broad spatiotemporal scales. However, sustained long-term monitoring poses significant challenges, particularly in regions. In this context, ecoacoustics emerges as a promising approach to address gap. Yet, harnessing insights from extensive acoustic datasets presents its own set such time and expertise needed label species information recordings. Here, study an investigating soundscapes: use deep neural network trained time-of-day estimation. This research endeavors (1) provide qualitative analysis temporal variation (daily monthly) soundscape using conventional ecoacoustic indices embeddings, (2) compare predictive power both methods for estimation, (3) performance supervised classification unsupervised clustering specific recording site, habitat type, season. The study's findings reveal that proposed embeddings exhibit overall comparable performance. article concludes by discussing potential avenues further refinement method, which will contribute understanding across space.

Language: Английский

Citations

0
Flexible and Interpretable Soundscape Analysis for Biodiversity Assessment and Ecosystem Health for Domain Experts DOI
Rida Saghir

Published: March 18, 2025

Language: Английский

Citations

0
EcoScape Analyzer: A Tool for Performing Soundscape Analysis With Flexible Pipeline for Biodiversity Assessment DOI
Rida Saghir, Ivan Braga Campos, Thiago S. Gouvêa

et al.

Published: March 18, 2025

Language: Английский

Citations

0
Meta-Embedded Clustering (MEC): A new method for improving clustering quality in unlabeled bird sound datasets DOI Creative Commons
Joachim POUTARAUD, Jérôme Sueur,

Christophe Thébaud

et al.

Ecological Informatics, Journal Year: 2024, Volume and Issue: 82, P. 102687 - 102687

Published: June 15, 2024

In recent years, ecoacoustics has offered an alternative to traditional biodiversity monitoring techniques with the development of passive acoustic (PAM) systems allowing, among others, detect and identify species that are difficult by human observers, automatically. PAM typically generate large audio datasets, but using these infer ecologically meaningful information remains challenging. most cases, several thousand hours recordings need be manually labeled experts limiting operability systems. Based on developments meta-learning algorithms unsupervised learning techniques, we propose here Meta-Embedded Clustering (MEC), a new method high potential for improving clustering quality in unlabeled bird sound datasets. MEC is organized two main steps, with: (a) fine-tuning pretrained convolutional neural network (CNN) backbone different pseudo-labeled data, (b) manually-labeled sounds latent space based vector embeddings extracted from fine-tuned CNN. The significantly enhanced average performance less than 1% more 80%, greatly outperforming approach relying solely CNN features general neotropical database. However, this came cost excluding portion data categorized as noise. By should facilitate work ecoacousticians managing units song/call clustered according their similarities, identifying clusters undetected approaches.

Language: Английский

Citations

0