Large-scale unsupervised clustering of Orca vocalizations: a model for describing orca communication systems

Poupard, Marion; Best, Paul; Schlüter, Jan; Symonds, Helena; Spong, Paul; Glotin, Hervé

doi:10.7287/peerj.preprints.27979v1

Cited by 5 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The strengths of CD lie primarily in its high speed and diverse application possibilities with high-dimensional data [ 41 , 42 ]. In addition, a new hierarchical density-based clustering algorithm (HDBSCAN), introduced by Campello et al in 2013 [ 43 ], will be applied and compared, which was first used in bioacoustics in 2020 [ 44 , 45 ]. This algorithm has the advantage that outliers are recognized as noise and discarded, so data points that are difficult to assign are ignored and do not negatively affect the clustering result.…”

Section: Introductionmentioning

confidence: 99%

Introducing the Software CASE (Cluster and Analyze Sound Events) by Comparing Different Clustering Methods and Audio Transformation Techniques Using Animal Vocalizations

Schneider

Hammerschmidt

Dierkes

2022

Animals

View full text Add to dashboard Cite

Unsupervised clustering algorithms are widely used in ecology and conservation to classify animal sounds, but also offer several advantages in basic bioacoustics research. Consequently, it is important to overcome the existing challenges. A common practice is extracting the acoustic features of vocalizations one-dimensionally, only extracting an average value for a given feature for the entire vocalization. With frequency-modulated vocalizations, whose acoustic features can change over time, this can lead to insufficient characterization. Whether the necessary parameters have been set correctly and the obtained clustering result reliably classifies the vocalizations subsequently often remains unclear. The presented software, CASE, is intended to overcome these challenges. Established and new unsupervised clustering methods (community detection, affinity propagation, HDBSCAN, and fuzzy clustering) are tested in combination with various classifiers (k-nearest neighbor, dynamic time-warping, and cross-correlation) using differently transformed animal vocalizations. These methods are compared with predefined clusters to determine their strengths and weaknesses. In addition, a multidimensional data transformation procedure is presented that better represents the course of multiple acoustic features. The results suggest that, especially with frequency-modulated vocalizations, clustering is more applicable with multidimensional feature extraction compared with one-dimensional feature extraction. The characterization and clustering of vocalizations in multidimensional space offer great potential for future bioacoustic studies. The software CASE includes the developed method of multidimensional feature extraction, as well as all used clustering methods. It allows quickly applying several clustering algorithms to one data set to compare their results and to verify their reliability based on their consistency. Moreover, the software CASE determines the optimal values of most of the necessary parameters automatically. To take advantage of these benefits, the software CASE is provided for free download.

show abstract

Section: Introductionmentioning

confidence: 99%

Introducing the Software CASE (Cluster and Analyze Sound Events) by Comparing Different Clustering Methods and Audio Transformation Techniques Using Animal Vocalizations

Schneider

Hammerschmidt

Dierkes

2022

Animals

View full text Add to dashboard Cite

show abstract

“…Factors such as environmental noise can pose limitations to the automatic approach, potentially leading to biased results (Brumm et al, 2017). Neural networks and deep learning have already been used to automatically detect orca sound samples for species identification (Poupard et al, 2019a) and classification of calls (Bergler et al, 2019; Poupard et al, 2019b). Dynamic type warping (DTW), a method for measuring pairwise acoustic similarities based on manually extracted fundamental frequency contours, has been used previously for orca sound type classification in both captivity and natural conditions (Brown et al, 2006; Brown & Miller, 2007).…”

Section: Introductionmentioning

confidence: 99%

Towards automatic detection and classification of orca (Orcinus orca) calls using cross‐correlation methods

Palmero

Guidi

Kulikovskiy

et al. 2022

Marine Mammal Science

View full text Add to dashboard Cite

Orcas (Orcinus orca) are known for complex vocalization. Their social structure consists of pods with unique call repertoires and clans sharing vocal traditions. Call repertoires are typically established visually and aurally and are used for pod identification. Automatic tools are, however, more suited for large data sets. An Icelandic orca pod occurring in 2019 in the Ligurian Sea provided a unique occasion for collecting recordings of an isolated pod in natural conditions.Recordings were analyzed visually and aurally to create a pod catalog. The R package "warbleR" was used for the first time on a small subsample of orca data to automatically detect sound samples and classify sound types. We found cross-correlation methods with the Pearson correlation coefficient (PCC) to successfully classify sound types, though false positives occur. We compared our catalog to Icelandic and Antarctic ones checking for similarities and dissimilarities. We found five matches in the Icelandic catalog, two of which had high PCCs ranges (0.62-0.67; 0.60-0.65). Our automatic approach was limited by background noise and variability of orca vocalizations, and it was computationally demanding. We show cross-correlation

show abstract

“…Neural networks and deep learning have already been used to automatically detect orca sounds with the purpose of species identi cation (Poupard et al, 2019a. ) and call classi cation (Bergler et al, 2019;Poupard et al, 2019b). Recent studies used dynamic type warping (DTW), a method for measuring pairwise acoustic dissimilarity based on dominant or fundamental frequency contours, to replace human classi cation and classi ed orca sounds in both captivity and natural conditions (Brown et al, 2006;Brown & Miller, 2007).…”

Section: Introductionmentioning

confidence: 99%

Towards automatic detection and classification of orca (Orcinus orca) calls using cross-correlation methods

Palmero

Guidi

Kulikovskiy

et al. 2021

Preprint

View full text Add to dashboard Cite

Orca (Orcinus orca) is known for complex vocalisation. Their social structure consists of pods and clans sharing unique dialects due to geographic isolation. Sound type repertoires are fundamental for monitoring orca populations and are typically created visually and aurally. An orca pod occurring in the Ligurian Sea (Pelagos Sanctuary) in December 2019 provided a unique occasion for long-term recordings. The numerous data collected with the bottom recorder were analysed with a traditional human-driven inspection to create a repertoire of this pod and to compare it to catalogues from different orca populations (Icelandic and Antarctic) investigating its origins. Automatic signal detection and cross-correlation methods (R package warbleR) were used for the first time in orca studies. We found the Pearson cross-correlation method to be efficient for most pairwise calculations (> 85%) but with false positives. One sound type from our repertoire presented a high positive match (range 0.62–0.67) with one from the Icelandic catalogue, which was confirmed visually and aurally. Our first attempt to automatically classify orca sound types presented limitations due to background noise and sound complexity of orca communication. We show cross-correlation methods can be a powerful tool for sound type classification in combination with conventional methods.

show abstract

Large-scale unsupervised clustering of Orca vocalizations: a model for describing orca communication systems

Cited by 5 publications

References 21 publications

Introducing the Software CASE (Cluster and Analyze Sound Events) by Comparing Different Clustering Methods and Audio Transformation Techniques Using Animal Vocalizations

Introducing the Software CASE (Cluster and Analyze Sound Events) by Comparing Different Clustering Methods and Audio Transformation Techniques Using Animal Vocalizations

Towards automatic detection and classification of orca (Orcinus orca) calls using cross‐correlation methods

Towards automatic detection and classification of orca (Orcinus orca) calls using cross-correlation methods

Contact Info

Product

Resources

About