Testing the temporal robustness of an automatic aural classifier.

Murphy, Stefan M.; Hines, Paul C.

doi:10.1121/1.3385381

Cited by 3 publications

(9 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Also of note is that the elements above the main diagonal are darker (i.e., represent better performance) than those below the main diagonal. A similar pattern was noted in the SNR-dependence investigation of Murphy and Hines [7], suggesting that SNR may, at least in part, be driving the decrease in classifier performance. Mouy et al [14] also noted that false negative rates increased as SNR decreased.…”

Section: Figure 3: Aural Classifier Performance Matrices Generated Frsupporting

confidence: 71%

“…Alternatively, it may be found that many of the perceptual features are environment-sensitive and therefore it is unreasonable to exclude all of them. In this case, a strategy may instead be developed to generate training sets for the classifier that take propagation-related signal distortion into account; this could be done either by acoustically distorting the training data by propagating them through a modeled environment or by including vocalizations in the training set that came from a variety of propagation ranges as done in [7,8].…”

mentioning

confidence: 99%

See 1 more Smart Citation

Quantifying the Effects of Propagation on Classification of Cetacean Vocalizations

Hines¹,

Binder²

2014

Self Cite

View full text Add to dashboard Cite

To develop a robust automatic classifier with a high probability of detection and a low false alarm rate that can classify vocalizations from a variety of cetacean species in diverse ocean environments. OBJECTIVES In previous work as part of ONR grant N000141210139 a unique automatic classifier developed by the PI that uses perceptual signal features-features similar to those employed by the human auditory system-was employed to successfully classify anthropogenic transients, and vocalizations from five cetacean species. Although this is a significant achievement, successful implementation of this (or any) classifier requires that it be temporally and spatially robust. The primary goal will be to address the question: "Will it work on vocalization data from these species collected under different environmental conditions?" To examine this, discriminant analysis will be used to rank the aural features in terms of their ability to separate the vocalizations between species. Then, the more highly ranked features will be tested for robustness. This will be done by performing a propagation experiment using cetacean vocalizations and synthetically generated calls as source signals, and testing the received signals with the classifier. The measurements will be complemented by comparing experimental results to propagation model results with the goal of generalizing the results to other ocean environments. APPROACH The research is part of a PhD program undertaken by Ms. Carolyn Binder under the supervision of Dr. Paul C. Hines. The postgraduate program is being conducted collaboratively in the Oceanography and Report Documentation Page Form Approved OMB No. 0704-0188 Public reporting burden for the 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 this burden, to Washington Headquarters Services, Directorate for Information Operations and Reports, 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 a penalty for failing to comply with a collection of information if it does not display a currently valid OMB control number.

show abstract

Section: Figure 3: Aural Classifier Performance Matrices Generated Frsupporting

confidence: 71%

mentioning

confidence: 99%

Quantifying the Effects of Propagation on Classification of Cetacean Vocalizations

Hines¹,

Binder²

2014

Self Cite

View full text Add to dashboard Cite

show abstract

“…To In general, the performance matrices generated using data from each recording unit displayed similar trends -the example results in Hines [79], further suggesting that SNR, at least in part, was responsible for driving the observed decrease in classifier performance. Mouy et al [23] also noted that false negative rates increased as SNR decreased.…”

Section: Training Set Selectionmentioning

confidence: 63%

“…In this section, we consider if high SNR signals recorded close to the sound source are the best for training the aural classifier. Murphy and Hines [79] concluded in their research with active sonar echoes that given two or more datasets with different SNR, the aural classifier performed better across multiple SNR regimes when the classifier was trained on a low SNR dataset. They ascribed this finding to the perceptual feature subset selected based on the signals contained in the training set -features selected from a high SNR training set did not perform well at low SNRs because the signal characteristics these features depended on became lost in the noise as SNR decreased, whereas features that were selected from the low SNR signals relied on signal characteristics that were still present in the higher SNR signals.…”

Section: Training Set Selectionmentioning

confidence: 99%

Impacts of environment-dependent acoustic propagation on passive acoustic monitoring of cetaceans

Binder

Hines

2016

Journal of the Acoustical Society of America

Self Cite

View full text Add to dashboard Cite

Significant effort has been made over the last few decades to develop automated passive acoustic monitoring (PAM) systems capable of classifying cetaceans at the species level; however, these systems often require tuning when deployed in different environments. Anecdotal evidence suggests that this requirement to adjust a PAM system's parameters is partially due to differences in the acoustic propagation characteristics. The environmentdependent propagation characteristics create variation in how a cetacean vocalization is distorted after it is emitted. If these difference are not accounted for it could reduce the performance of automated PAM systems. An aural classifier developed at Defence R&D Canada (DRDC) has been used successfully for inter-species discrimination of cetaceans. Accurate results are obtained by using perceptual signal features that model the features employed by the human auditory system. In this thesis, a combination of an at-sea experiment and simulations with modified bowhead and humpback whale vocalizations was conducted to investigate the robustness of the classifier performance to signal distortion as a function of propagation range. It was found that in many environments classification performance degraded with increasing range, largely due to decreased signal-to-noise ratio (SNR); however, in some environments as much as 40 % of the performance reduction was attributed to signal distortion resulting from environment-dependent propagation. It was found that sound speed profiles resulting in considerable boundary interaction were important for producing sufficient signal distortion to affect PAM performance, relative to the impacts of SNR. Therefore, in some environments the ocean acoustic properties should be taken into account when characterizing performance of automated PAM systems. For the environments in which signal-to-noise issues dominate, the use of multi-element arrays is expected to increase the performance of automated recognition systems beyond the minor improvements to be gained from adjusting a PAM system's parameters. Nonetheless, propagation modelling should be used to complement PAM experiments to account for bias in probability of detection estimates resulting from environment-dependent acoustic propagation.

show abstract

“…Propagation effects on the performance of the aural classifier may be analogous to the dependence of the aural classifier's performance on the Signal to Noise Ratio (SNR) levels of active sonar echoes. Work done by Murphy and Hines [7] demonstrated that higher performance was achieved when the classifier was trained with echoes of a similar SNR to the SNR of the data in a test set. It was also found that if echoes in the test set had a range of SNR values, the best performance was achieved by training the classifier with echoes that also encompassed a wide range of SNR values.…”

Section: Resultsmentioning

confidence: 99%

Robustness of perceptual features used for automatic aural classification to propagation effects

Binder

Hines

Pecknold

et al. 2013

The Journal of the Acoustical Society of America

Self Cite

View full text Add to dashboard Cite

Previous effort has shown that a prototype aural classifier developed at Defence R&D Canada can be used to reduce false alarm rates and successfully discriminate cetacean vocalizations from several species. The aural classifier achieves accurate results by using perceptual signal features that model the features employed by the human auditory system. Current work focuses on determining the robustness of the perceptual features to propagation effects for two of the cetacean species studied previously-bowhead and humpback whales. To this end, classification results are compared for the original vocalizations to classification results obtained after the vocalizations were re-transmitted underwater over ranges of 2 to 10 km. Additional insight into the propagation effects is gained from transmission of synthetic bowhead and humpback vocalizations, designed to have features similar to the most important aural features for classification of bowhead and humpback vocalizations. Each perceptual feature is examined individually to determine its robustness to propagation effects compared to the other aural features. To gain further understanding of propagation effects on the features, preliminary propagation modelling results are presented in addition to experimental data.

show abstract

Testing the temporal robustness of an automatic aural classifier.

Cited by 3 publications

References 0 publications

Quantifying the Effects of Propagation on Classification of Cetacean Vocalizations

Quantifying the Effects of Propagation on Classification of Cetacean Vocalizations

Impacts of environment-dependent acoustic propagation on passive acoustic monitoring of cetaceans

Robustness of perceptual features used for automatic aural classification to propagation effects

Contact Info

Product

Resources

About