2017
DOI: 10.1111/rssc.12217
|View full text |Cite
|
Sign up to set email alerts
|

Bat Echolocation Call Identification for Biodiversity Monitoring: A Probabilistic Approach

Abstract: Summary. Bat echolocation call identification methods are important in developing efficient cost-effective methods for large-scale bioacoustic surveys for global biodiversity monitoring and conservation planning. Such methods need to provide interpretable probabilistic predictions of species since they will be applied across many different taxa in a diverse set of applications and environments. We develop such a method using a multinomial probit likelihood with independent Gaussian process priors and study its… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
4
1

Citation Types

0
12
0
1

Year Published

2017
2017
2024
2024

Publication Types

Select...
7
2
1

Relationship

3
7

Authors

Journals

citations
Cited by 18 publications
(13 citation statements)
references
References 49 publications
0
12
0
1
Order By: Relevance
“…Existing methods typically extract a set of audio features (such as call duration, mean frequency, and mean amplitude) from high quality search-phase echolocation call reference libraries to train machine learning algorithms to classify unknown calls to species [ 11 , 14 19 ]. Instead of using manually defined features, another set of approaches attempt to learn representation directly from spectrograms [ 20 , 21 ]. Localising audio events in time (defined here as ‘detection’), is an important challenge in itself, and is often a necessary pre-processing step for species classification [ 22 ].…”
Section: Introductionmentioning
confidence: 99%
“…Existing methods typically extract a set of audio features (such as call duration, mean frequency, and mean amplitude) from high quality search-phase echolocation call reference libraries to train machine learning algorithms to classify unknown calls to species [ 11 , 14 19 ]. Instead of using manually defined features, another set of approaches attempt to learn representation directly from spectrograms [ 20 , 21 ]. Localising audio events in time (defined here as ‘detection’), is an important challenge in itself, and is often a necessary pre-processing step for species classification [ 22 ].…”
Section: Introductionmentioning
confidence: 99%
“…Looking forward, several emerging methods are substantially improving detection and classification accuracies by learning representations from spectrogram data, such as unsupervised feature extraction (Salamon & Bello, 2015;Stowell & Plumbley, 2014) and dynamic time warping based feature representations (Stathopoulos, Zamora-Gutierrez, Jones, & Girolami, 2017). Deep convolutional neural networks (CNNs) are particularly promising, since these can learn discriminating spectro-temporal information directly from annotated spectrograms (bypassing a separate feature extraction stage), improving their robustness to sound overlap and caller distance (Goeau et al, 2016) (Figure 4d).…”
Section: Emerging Innovations In Sound Identificationmentioning
confidence: 99%
“…Sensor technologies are used for tracking people and animals in urban and rural spaces [48,53]. The data collected are used to analyse movement patterns, people counts, etc., to better understand population and conservation concerns [48,50,53]. The presence and movement of humans in a given place and time is typically detected by recording sounds, movements or radio/WiFi signals emitted from a device they are carrying.…”
Section: Introductionmentioning
confidence: 99%