Deep Learning for Ethoacoustics of Oreas on three years pentaphonie continuous recording at Orealab revealing tide, moon and diel effects

Poupard, Marion; Best, Paul; Schlüter, Jan; Prévot, Jean‐Marc; Symonds, Helena; Spong, Paul; Glotin, Hervé

doi:10.1109/oceanse.2019.8867251

Cited by 8 publications

(9 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To accomplish this, we analyzed three different natural scenes or observations. Previous shore-based fixed hydrophone systems enabled the detection and tracking of orcas from the same population (Grebner, 2009;Bergler et al, 2019;Poupard et al, 2019a;Best et al, 2020) but did not succeed in associating individuals with calls or determine precise individual pattern variations of communication between groups. Tracking of a few individual orcas in another population, without visual checking, has been realized in a complex pilot study using 14 hydrophones deployed in three compact arrays (Gassmann et al, 2013).…”

Section: Discussionmentioning

confidence: 99%

“…Once the passage was finished, the whole team confirmed the presence of the individuals using the photographs and the photo-identification catalog (Towers, 2015). The map of viewing and listening ranges is available in Supplementary Figure 3 The listening range was estimated in consultation with the Orcalab experts H.S and P.S by means of the hydrophone array on site 24 h a day (Poupard et al, 2019a).…”

Section: Visual Data Acquisitionmentioning

confidence: 99%

See 1 more Smart Citation

Intra-Group Orca Call Rate Modulation Estimation Using Compact Four Hydrophones Array

Poupard

Symonds²,

Spong³

et al. 2021

Front. Mar. Sci.

Self Cite

View full text Add to dashboard Cite

Acoustic emissions are vital for orca (Orcinus orca) socializing, hunting, and maintaing spatial awareness. Studying the acoustic emissions of orcas on an individual basis often results in interference with their natural behaviors through mounting tags or following by boat. In order to analyze their inter- and intra-group communication, we propose a study allowing us to associate vocalizations with their emitter (matriline and when possible individual). Such a non-interfering device for allocating calls to individual orcas could substantially boost our understanding of their complex acoustic world. Our experimental protocol was based on a compact array of four hydrophones fixed near the shore, operable up to 1 km away from the path of orcas. It was used during summer 2019 at the research station OrcaLab, northern Vancouver Island, Canada. A total of 722 calls were extracted, jointly with visual identification and azimuth of surfacing orcas, allowing validation of the acoustic diarization and azimuth estimations of the orca calls. We then calculated the Call Rate (CR) for each matriline or when possible individual in order to describe their acoustic activity. Preliminary results show that CR could be modulated according to the distance of the signaler from a group, the presence of another group, or anthropic pressure.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Visual Data Acquisitionmentioning

confidence: 99%

Intra-Group Orca Call Rate Modulation Estimation Using Compact Four Hydrophones Array

Poupard

Symonds²,

Spong³

et al. 2021

Front. Mar. Sci.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Our shore based study looked at orca communication in the wild at a unique level of precision. Other shore based fixed hydrophone systems enabled the detection and tracking of orcas from the same population [36][37][38] but did not succeed in associating individuals with calls, or determine precise individual pattern variations of communication between groups. This work does not pretend to provide all explanations for the type of calls, but rather the CR.…”

Section: Discussionmentioning

confidence: 99%

Evidences of Intra-Group Orca Call Rate Modulation Using A Small-Aperture Four Hydrophone Array

Poupard¹,

Symonds²,

Spong³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Acoustic emissions are vital to orcas (Orcinus orca) to socialize, hunt, orient, and maintain spatial awareness. In order to better analyze their inter and intra-group communication, we propose a novel protocol that allows us to associate vocalizations with their emitter (individual/matriline). Our approach is based on a low cost small-aperture four hydrophone array fixed near the shore up to a few km away from the orcas’ path, operated in conjunction with visual identification. It was conducted in the summer of 2019 off northern Vancouver Island, Canada, at the research station OrcaLab. A total of 722 calls were extracted and localized in azimuth via the hydrophone array from 3 case studies in which different events took place.We then calculated the Call Rate (CR) for each individual/matriline in order to describe their acoustic activity. Results show that CR is modulated according to the distance of the signaler from the joint group, the presence of another group, and the anthropic pressure (nearby cruise ship). This shows evidence of intertwined calls. This protocol does not interfere with the animals and opens new perspectives towards inter and intra-group communication analysis.

show abstract

“…Bioacousticians, that often have to deal with complex sound recognition tasks themselves (e.g from species to acoustic unit classification) are now taking hold of this method. Deep learning has already been successfully applied to tasks such as bird classification [1], or orca vocalization detection [2]- [4]. However bioacousticians that use deep learning often still have to cope with the lack of broad-domain clean datasets.…”

Section: Introduction a Deep Learnging And Bioacousticsmentioning

confidence: 99%

Deep Learning and Domain Transfer for Orca Vocalization Detection

Best

Ferrari

Poupard

et al. 2020

2020 International Joint Conference on Neural Networks (IJCNN)

Self Cite

View full text Add to dashboard Cite

In this paper, we study the difficulties of domain transfer when training deep learning models, on a specific task that is orca vocalization detection. Deep learning appears to be an answer to many sound recognition tasks in human speech analysis as well as in bioacoustics. This method allows to learn from large amounts of data, and find the best scoring way to discriminate between classes (e.g. orca vocalization and other sounds). However, to learn the perfect data representation and discrimination boundaries, all possible data configurations need to be processed. This causes problems when those configurations are ever changing (e.g. in our experiment, a change in the recording system happened to considerably disturb our previously well performing model). We thus explore approaches to compensate on the difficulties faced with domain transfer, with two convolutionnal neural networks (CNN) architectures, one that works in the time-frequency domain, and one that works directly on the time domain.

show abstract

Deep Learning for Ethoacoustics of Oreas on three years pentaphonie continuous recording at Orealab revealing tide, moon and diel effects

Cited by 8 publications

References 23 publications

Intra-Group Orca Call Rate Modulation Estimation Using Compact Four Hydrophones Array

Intra-Group Orca Call Rate Modulation Estimation Using Compact Four Hydrophones Array

Evidences of Intra-Group Orca Call Rate Modulation Using A Small-Aperture Four Hydrophone Array

Deep Learning and Domain Transfer for Orca Vocalization Detection

Contact Info

Product

Resources

About