How Can Dolphins Recognize Fish According to Their Echoes? A Statistical Analysis of Fish Echoes

Yovel, Yossi; Au, Whitlow W. L.

doi:10.1371/journal.pone.0014054

Cited by 15 publications

(18 citation statements)

References 17 publications

(23 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…For dolphins, artificial neural networks have been applied to model dolphin sonar, specifically for discriminating differences in the wall thickness of cylinders using time and frequency information from the echoes (Au et al 1995). Also, support vector machines and quadratic discriminant function analysis have been used to classify fish species according to their echoes using a dolphin-emulating sonar system (Yovel and Au 2010), and Gaussian mixture models and support vector machines have been employed to classify echolocation clicks from three species of odontocetes (Roch et al 2008). Differentiation of categories or graded barks in mother-calf vocal communication in Atlantic walrus have been analyzed with artificial neural networks and discriminant functions (Charrier et al 2010).…”

Section: Introductionmentioning

confidence: 99%

Comparing supervised learning methods for classifying sex, age, context and individual Mudi dogs from barking

et al. 2014

View full text Add to dashboard Cite

Barking is perhaps the most characteristic form of vocalization in dogs; however, very little is known about its role in the intraspecific communication of this species. Besides the obvious need for ethological research, both in the field and in the laboratory, the possible information content of barks can also be explored by computerized acoustic analyses. This study compares four different supervised learning methods (naive Bayes, classification trees, k-nearest neighbors and logistic regression) combined with three strategies for selecting variables (all variables, filter and wrapper feature subset selections) to classify Mudi dogs by sex, age, context and individual from their barks. The classification accuracy of the models obtained was estimated by means of K-fold cross-validation. Percentages of correct classifications were 85.13 % for determining sex, 80.25 % for predicting age (recodified as young, adult and old), 55.50 % for classifying contexts (seven situations) and 67.63 % for recognizing individuals (8 dogs), so the results are encouraging. The best-performing method was k-nearest neighbors following a A. Larranaga Student at the Universidad Alfonso X El Sabio, Av. Universidad, 1, 28691 Villanueva de la Canada, Madrid, Spain wrapper feature selection approach. The results for classifying contexts and recognizing individual dogs were better with this method than they were for other approaches reported in the specialized literature. This is the first time that the sex and age of domestic dogs have been predicted with the help of sound analysis. This study shows that dog barks carry ample information regarding the caller's indexical features. Our computerized analysis provides indirect proof that barks may serve as an important source of information for dogs as well.

show abstract

Section: Introductionmentioning

confidence: 99%

Comparing supervised learning methods for classifying sex, age, context and individual Mudi dogs from barking

et al. 2014

View full text Add to dashboard Cite

show abstract

“…This raises the question of whether the results from psychophysical echolocation trials made on bottlenose dolphins in captivity are directly applicable when trying to understand the biosonar performance of free-ranging animals. In fact, the properties of captive T. truncatus clicks have been used to estimate the echolocation performance in terms of detection ranges (Au et al, 2007; and target discrimination abilities in the wild (Au et al, 2009;Yovel and Au, 2010). Although such ecophysiological inferences are important for understanding the evolution and use of biosonar in the wild, they critically hinge on the fact that the chosen assumptions (out of many) from the captive T. truncatus are representative for their wild conspecifics.…”

mentioning

confidence: 99%

Source parameters of echolocation clicks from wild bottlenose dolphins (Tursiops aduncusandTursiops truncatus)

Wahlberg

Jensen

Soto

et al. 2011

The Journal of the Acoustical Society of America

View full text Add to dashboard Cite

The Indian Ocean and Atlantic bottlenose dolphins (Tursiops aduncus and Tursiops truncatus) are among the best studied echolocating toothed whales. However, almost all echolocation studies on bottlenose dolphins have been made with captive animals, and the echolocation signals of freeranging animals have not been quantified. Here, biosonar source parameters from wild T. aduncus and T. truncatus were measured with linear three-and four-hydrophone arrays in four geographic locations. The two species had similar source parameters, with source levels of 177-228 dB re 1 lPa peak to peak, click durations of 8-72 ls, centroid frequencies of 33-109 kHz and rms bandwidths between 23 and 54 kHz. T. aduncus clicks had a higher frequency emphasis than T. truncatus. The transmission directionality index was up to 3 dB higher for T. aduncus (29 dB) as compared to T. truncatus (26 dB). The high directionality of T. aduncus does not appear to be only a physical consequence of a higher frequency emphasis in clicks, but may also be caused by differences in the internal properties of the sound production system.

show abstract

“…A pulse is transmitted into the medium where the object of interest is located. The sonar receiver receives acoustical reflections of the transmitted pulse from the medium, which convey information about object location, structure, and sometimes composition [21]. Fig.…”

Section: A Sonar Signal and Propagation Modelingmentioning

confidence: 99%

Noncontact Wideband Sonar for Human Activity Detection and Classification

2014

Self Cite

View full text Add to dashboard Cite

This paper suggests using a wideband sonar system to detect and classify human activity in indoor environment. The high bandwidth enables precise tracking of body parts, and its enhanced correlation properties can be used to distinguish between human and nonhuman objects. Maximal Likelihood (ML) criterions to derive kinematic features and processing methods to estimate the subject activity level and activity type were derived and tailored to the wideband sonar. For tracking and association of the echoes reflected from the different body part, an efficient approximation of the sequential ML estimator was derived in the natural time-space domain, which eases the exploitation of the a priori knowledge about the human subject target. For classification of the activity, a weighted two level nested k-nearest neighbor classifier was applied on only four kinematic features. A set of experiments with five subjects, performing three different activity types of standing, walking, and swinging upper limbs, was carried out in a typical indoor environment. The proposed technology has managed to classify well the different activity types and demonstrated the potential of this technology for continuous assessment of various kinematic features of humans in indoor environment with reduced costs, under any light, smoke, or humidity conditions. This can be useful for instance for monitoring patients at home, and for detecting intruders.

show abstract

How Can Dolphins Recognize Fish According to Their Echoes? A Statistical Analysis of Fish Echoes

Cited by 15 publications

References 17 publications

Comparing supervised learning methods for classifying sex, age, context and individual Mudi dogs from barking

Comparing supervised learning methods for classifying sex, age, context and individual Mudi dogs from barking

Source parameters of echolocation clicks from wild bottlenose dolphins (Tursiops aduncusandTursiops truncatus)

Noncontact Wideband Sonar for Human Activity Detection and Classification

Contact Info

Product

Resources

About