Classification of Transient Sonar Sounds Using Perceptually Motivated Features

Tucker, Simon; Brown, Guy J.

doi:10.1109/joe.2005.850910

Cited by 47 publications

(40 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, we consider for the classification of volcano-seismic data, an extensive set of features proposed in several fields such as seismic [2], acoustics (environmental, bio-acoustics, animal [3], [4], [5], [6] or anthropic [7] ambient and/or landscape noises), speech and speech analysis [8], [9], [10] and music signals [11]. More in details, we take into account features such as those proposed in the recent work [2] where seismic waves are represented by few classically features such as duration time, statistical descriptors (skewness, kurtosis, statistics ratios) and fundamental frequencies.…”

Section: Related Workmentioning

confidence: 99%

“…More in details, we take into account features such as those proposed in the recent work [2] where seismic waves are represented by few classically features such as duration time, statistical descriptors (skewness, kurtosis, statistics ratios) and fundamental frequencies. In classification of transient sonar sounds, the work in [7] considers more than 20 features. Those features are mainly composed of descriptors of the signal shape (rate of attack and decay, time of the main peak), statistical moments (mean skewness and kurtosis) and of signal power (peak power, average and power standard deviation).…”

Section: Related Workmentioning

confidence: 99%

“…[7], [2] closely monitored especially by the Instituto Geofisico del Péru. In this study, we used data recorded over 6 years (between 2000 and 2006) by a 1 Hz vertical sensor at station UB1 with a sampling rate of 100Hz and transmitted in real time.…”

Section: Featurementioning

confidence: 99%

See 2 more Smart Citations

Machine learning for automatic classification of volcano-seismic signatures

Malfante¹,

Mura²,

Mars³

et al. 2017

2017 25th European Signal Processing Conference (EUSIPCO)

View full text Add to dashboard Cite

Abstract-The evaluation and prediction of volcanoes activities and associated risks is still a timely and open issue. The amount of volcano-seismic data acquired by recent monitoring stations is huge (e.g., several years of continuous recordings), thereby making machine learning absolutely necessary for their automatic analysis. The transient nature of the volcano-seismic signatures of interest further enforces the need of automatic detection and classification of such events. In this paper, we present a novel architecture for automatic classification of volcano-seismic events based on a comprehensive signal representation with a large feature set. To the best of our knowledge this is one of the first attempts to automatize the classification task of these signals. The proposed approach relies on supervised machine learning techniques to build a prediction model.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Machine learning for automatic classification of volcano-seismic signatures

Malfante¹,

Mura²,

Mars³

et al. 2017

2017 25th European Signal Processing Conference (EUSIPCO)

View full text Add to dashboard Cite

show abstract

“…As timbre is most likely a multidimensional attribute, most studies in timbre rely on a numerical technique known as multidimensional scaling (MDS) [1,[10][11][12][13][14][15]. MDS is a nonlinear data analysis technique which takes known (perceptual) distances between data points and identifies a low dimensional Euclidian space that maintains those distances [16,17].…”

Section: Aural Similaritymentioning

confidence: 99%

“…In previous studies [10,12,14,15], a number of hypothesized signal features, (. ), are first calculated for each sound.…”

Section: Standard Mds Analysismentioning

confidence: 99%

Perceptually-Driven Signal Analysis for Acoustic Event Classification

McLaughlin¹,

Philips²,

Pitton³

2007

View full text Add to dashboard Cite

The public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources. gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing the burden, to Department of Defense, Washington Headquarters Services, Directorate for Information Operations and Reports (0704-0188), 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302. Respondents should be aware that notwithstanding any other provision of law, no person shall be subject to any penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number. PLEASE DO NOT RETURN YOUR FORM TO THE ABOVE ADDRESS. REPORT DATE (DD-MM-YYYY)2. REPORT TYPE 3. DATES COVERED (From -To) SPONSORING/MONITORING AGENCY NAME(S) AND ADDRESSIES) 10. SPONSOR/MONITOR'S ACRONYM(S)Office of DISTRIBUTION/AVAILABILITY STATEMENTApproved for public release; distribution is unlimited. SUPPLEMENTARY NOTES ABSTRACTThis research develops a framework for employing perceptual information from human listening experiments to improve automatic event classification. We focus on the identification of new signal attributes, or features, that are able to predict the human performance observed in formal listening experiments. Using this framework, our newly identified features have the ability to elevate automatic classification performance closer to the level of human listeners.We develop several new methods for learning a perceptual feature transform from human similarity measures. We also develop a new approach for learning a perceptual distance metric. Our research demonstrates these new methods in the area of active sonar signal processing and confirms anecdotal evidence that human operators are adept in the task of discriminating between active sonar target and clutter echoes. We identify perceptual features and distance metrics using our novel methods. The results show better agreement with human performance than previous approaches. SUBJECT TERMS SECURITY CLASSIFICATION OF:17 20080211154 AbstractIn many acoustic signal processing applications human listeners are able to outperform automated processing techniques, particularly in the identification and classification of acoustic events. The research discussed in this paper develops a framework for employing perceptual information from human listening experiments to improve automatic event classification. We focus on the identification of new signal attributes, or features, that are able to predict the human performance observed in formal listening experiments. Using this framework, our newly identified features have the ability to elevate automatic classification performance closer to the level of human listeners.We develop several new methods for learning a perceptual feature transform from h...

show abstract

‘Posidonia meadows calling’: a ubiquitous fish sound with monitoring potential

Iorio

Raick

Parmentier

et al. 2018

Remote Sens Ecol Conserv

View full text Add to dashboard Cite

In the Mediterranean Sea, the seagrass Posidonia oceanica plays a key ecological role, and is protected by a range of legislation. Standard Posidonia monitoring programmes generally focus on the plant at different spatial and short temporal scales, without considering the organisms dependent on the ecosystem. Passive acoustic monitoring (PAM) has a high potential to non-intrusively monitor biological activities and biodiversity at high temporal resolution, and to assess ecosystem health. This is particularly relevant considering that Posidonia meadows host numerous sound-producing fish species. In this study, bottommoored hydrophones were deployed in nine Western Mediterranean meadows covering a distance of more than 200 km to identify acoustic features potentially relevant to monitor this critical habitat. Among eight identified fish sound categories, we found a single type of sound (that we will refer to as /kwa/) dominating the soundscape of Posidonia meadows over a time span of 7 months. Compared to other low-frequency sounds, the /kwa/ presented unique characteristics that suggest it is produced by a fish via fast contracting muscles. The /kwa/ was the only sound detectable under anthropogenic noise conditions, and little affected by it. Cluster analyses performed on 13 acoustic features revealed a high degree of call diversity. /Kwa/ diversity, combined with its large-scale (all meadows), long-term (7 months) occurrence and low noise interference, make the /kwa/ a promising candidate for PAM of Posidonia meadows. Furthermore, variability in acoustic features suggests a central role of the /kwa/ in communication. Overall, this work sets the basis for establishing the relevance of the /kwa/ in monitoring P. oceanica meadows and developing PAM techniques for this critical habitat.

show abstract

Classification of Transient Sonar Sounds Using Perceptually Motivated Features

Cited by 47 publications

References 22 publications

Machine learning for automatic classification of volcano-seismic signatures

Machine learning for automatic classification of volcano-seismic signatures

Perceptually-Driven Signal Analysis for Acoustic Event Classification

‘Posidonia meadows calling’: a ubiquitous fish sound with monitoring potential

Contact Info

Product

Resources

About