Species-specific audio detection: a comparison of three template-based detection algorithms using random forests

Bravo, Carlos J Corrada; Berríos, Rafael Álvarez; Aide, T. Mitchell

doi:10.7717/peerj-cs.113

Cited by 22 publications

(10 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Furthermore, from the results, the decision-tree method appears as one of the best classifiers in many temporally-aware methods. This fact is consistent with other studies where non-speech sounds (Pavlopoulos, Stasis & Loukis, 2004), or more specifically, environmental sounds (Bravo, Berríos & Aide, 2017) are considered.…”

Section: Discussionsupporting

confidence: 92%

Peer Review #2 of "Temporally-aware algorithms for the classification of anuran sounds (v0.1)"

2018

View full text Add to dashboard Cite

Several authors have shown that the sounds of anurans can be used as an indicator of climate change. Hence, the recording, storage and further processing of a huge number of anuran sounds, distributed over time and space, are required in order to obtain this indicator. Furthermore, it is desirable to have algorithms and tools for the automatic classification of the different classes of sounds. In this paper, six classification methods are proposed, all based on the data-mining domain, which strive to take advantage of the temporal character of the sounds. The definition and comparison of these classification methods is undertaken using several approaches. The main conclusions of this paper are that: i) the sliding window method attained the best results in the experiments presented, and even outperformed the Hidden Markov Models usually employed in similar applications; ii) noteworthy overall classification performance has been obtained, which is an especially striking result considering that the sounds analysed were affected by a highly noisy background; iii) the instance selection for the determination of the sounds in the training dataset offers better results than cross-validation techniques; and iv) the temporally-aware classifiers have revealed that they can obtain better performance than their non-temporally-aware counterparts.

show abstract

Section: Discussionsupporting

confidence: 92%

Peer Review #2 of "Temporally-aware algorithms for the classification of anuran sounds (v0.1)"

2018

View full text Add to dashboard Cite

show abstract

“…Several main concepts emerge from this work: first, although other auspicious classification methods implicitly strive to minimize false positives (e.g. Heinicke et al 2015;Bas et al 2017;Corrada-Bravo et al 2017;Ranjard et al 2017), none that we know of explicitly address false positive mitigation within the context of template-based or threshold-based detection. In addition to making binary predictions about each detection's class, this method also has the advantage of producing probability values for each detection, which may be aggregated in dynamic occupancy models to predict the overall probability of species occurrence (Balantic and Donovan 2019).…”

Section: Discussionmentioning

confidence: 99%

“…Regardless of the automated detection method employed, when acting without human assistance, computer-automated methods often produce an unacceptable number of false alarms, wherein non-target noise is detected and incorrectly assigned to a target species (Acevedo et al 2009). False alarm rates from computerautomated methods may vary widely from project to project based on the prevalence of similar sounds from non-target sources in the soundscape, acoustic characteristics of sounds made by the target species (Towsey et al 2012), the type of automated detection routine used (Corrada-Bravo et al 2017), the available number of target sound examples upon which automated methods may be trained (Stowell et al 2016), the quality of training data in terms of how well it represents the data that will be subject to the automated detection method (Knight and Bayne 2018), and selection of score thresholds above which detections may occur (Knight et al 2017).…”

Section: Introductionmentioning

confidence: 99%

Statistical learning mitigation of false positives from template-detected data in automated acoustic wildlife monitoring

Balantic

Donovan

2019

Bioacoustics

View full text Add to dashboard Cite

Audio sampling of the environment can provide long-term, landscape-scale presence-absence data to model populations of soundproducing wildlife. Automated detection systems allow researchers to avoid manually searching through large volumes of recordings, but often produce unacceptable false positive rates. We developed methods that allow researchers to improve template-based automated detection using a suite of statistical learning algorithms when false positive rates are problematic. To test our method, we acquired 668 hours of recordings in the Sonoran Desert, California USA between March 2016 and May 2017, and created spectrogram cross-correlation templates for three target avian species. We trained and tested five classification algorithms and four performance-weighted ensemble classifier methods on target signals and false alarms from March 2016, and then selected highperforming ensemble classifiers from the train/test phase to predict the class of new detections thereafter. For three target species, our ensemble classifiers were able to identify 98%, 81%, and 100% of false alarms compared with the baseline template detection system, and comparative positive predictive values improved from 6% to 69%, 87% to 95%, and 2% to 77%. We show that statistical learning approaches can be implemented to mitigate false detections acquired via template-based automated detection in automated acoustic wildlife monitoring.

show abstract

“…To analyze the occurrence of floodplain specialist species in the upstream sites, we used automated classification algorithms in the RFCx-ARBIMON platform to determine the presence or absence of 24 floodplain specialists (diurnal birds) in 93,435 audio recordings (between 05h00 to 18h00). Species-specific identification models allow the detection and analysis of target species in a large dataset and have been successfully used in several groups of organisms (Corrada-Bravo et al 2017;LeBien et al 2020).…”

Section: Analysis Of Bird Communitiesmentioning

confidence: 99%

Impacts of a large hydroelectric dam on the Madeira River (Brazil) on floodplain avifauna

et al. 2021

View full text Add to dashboard Cite

Hydroelectric dams represent an important threat to seasonally flooded environments in the Amazon basin. We aimed to evaluate how a dam in the Madeira River, one of the largest tributaries of the Amazonas River, affected floodplain avifauna. Bird occurrence was recorded through simultaneous passive acoustic monitoring in early successional vegetation and floodplain forest downstream from the dam and upstream in sites impacted by permanent flooding after dam reservoir filling. Species were identified through manual inspection and semi-automated classification of the recordings. To assess the similarity in vegetation between downstream and upstream sites, we used Landsat TM/ETM+ composite images from before (2009-2011) and after (2016-2018) reservoir filling. Downstream and upstream floodplain forest sites were similar before, but not after dam construction. Early successional vegetation sites were already different before dam construction. We recorded 195 bird species. While species richness did not differ between upstream and downstream sites, species composition differed significantly. Ten species were indicators of early successional vegetation upstream, and four downstream. Ten species were indicators of floodplain forest upstream, and 31 downstream. Seven of 24 floodplain specialist species were detected by the semi-automated classification only upstream. While we found some bird species characteristic of early successional vegetation in the upstream sites, we did not find most species characteristic of tall floodplain forest. Predominantly carnivorous, insectivorous, and nectarivorous species appear to have been replaced by generalist and widely distributed species.

show abstract

Species-specific audio detection: a comparison of three template-based detection algorithms using random forests

Cited by 22 publications

References 29 publications

Peer Review #2 of "Temporally-aware algorithms for the classification of anuran sounds (v0.1)"

Peer Review #2 of "Temporally-aware algorithms for the classification of anuran sounds (v0.1)"

Statistical learning mitigation of false positives from template-detected data in automated acoustic wildlife monitoring

Impacts of a large hydroelectric dam on the Madeira River (Brazil) on floodplain avifauna

Contact Info

Product

Resources

About