Review of Automatic Detection and Classification Techniques for Cetacean Vocalization

Abstract

Cetaceans have elicited the attention of researchers in recent decades due to their importance to the ecosystem and their economic values. They use sound for communication, echolocation and other social activities. Their sounds are highly non-stationary, transitory and range from short to long sounds. Passive acoustic monitoring (PAM) is a popular method used for monitoring cetaceans in their ecosystems. The volumes of data accumulated using PAM are usually big, so they are difficult to analyze using manual inspection. Therefore different techniques with mixed outcomes have been developed for the automatic detection and classification of signals of different cetacean species. So far, no single technique developed is perfect to detect and classify the vocalizations of over 82 known species due to variability in time-frequency, difference in the amplitude among species and within species' vocal repertoire, physical environment, among others. The accuracy of any detector or classifier depends on the technique adopted as well as the nature of the signal to be analyzed. In this article, we review the existing techniques for the automatic detection and classification of cetacean vocalizations. We categorize the surveyed techniques, while emphasizing the advantages and disadvantages of these techniques. The article suggests possible research directions that can improve existing detection and classification techniques. In addition, the article recommends other suitable techniques that can be used to analyze non-linear and non-stationary signals such as the cetaceans' signals. Several research have been dedicated to this topic, however, there is no review of these past results that gives a quick overview in the area of cetacean detection and classification. This review will help researchers and practitioners in the field to make insightful decisions based on their requirements.

Characterization and comparison of echolocation clicks of white-beaked dolphins (Lagenorhynchus albirostris) off the Northumberland coast, UK

“…Odontocetes continuously produce different types of pulsed sounds, which offer a suitable acoustic indicator for determining presence and behavior of individuals (Richardson et al, 1995;Temple et al, 2016). Characterization of clicks can facilitate the acoustic identification of species and is critical for future passive acoustic monitoring and research efforts (Wahlberg et al, 2011;Usman et al, 2020). Comparison of click train types used in different behavioral contexts may improve the understanding of the functional adaptions made by different echolocating species and give further insight into the drivers of acoustic behavior (Madsen and Surlykke, 2013).…”

Detection of Bryde's whale short pulse calls using time domain features with hidden Markov models

“…The MFCC typically uses a total number of cepstral coefficients n ranging from 10 to 14 [23], [32], which has been shown to produce low computational complexity compared to other HMM based feature extraction methods [53]. Here, the 12-dimensional MFCC-HMM and LPC-HMM models are used compared to the 3-dimensional PaMZ-HMM.…”

“…More so, the classification of cetacean signals is implemented using various optimal classifiers such as dynamic time warping (DTW) [16]- [18], artificial neural networks (ANN) [15], [19], support vector machine (SVM) [20], hidden Markov model (HMM) [21], [22], and so on. However, the choice of the detector and classifier, together with its performance depends on the feature vector, the species involved, the volume of data, and the location of recordings among others [23].…”

Improve automatic detection of animal call sequences with temporal context

“…DNNs have been used in the automatic recognition of vocalizations of birds [ 10 ], primates (e.g. [ 11 ]), marine mammals (see [ 12 ] for a review), marsupials (e.g. [ 13 ]), fishes (e.g.…”