Publications

Export 4 results:
Sort by: [ Author  (Asc)] Title Type Year
A B C D E F G H I J K L M N O P Q [R] S T U V W X Y Z   [Show ALL]
R
Roch, M. A., Soldevilla MS, Hoenigman R, Wiggins SM, Hildebrand J.  2008.  Comparison of machine learning techniques for the classification of echolocation clicks from three species of Odontocetes. Canadian Acoustics. 36:41-47. AbstractWebsite

A species detectorclassifier is presented which decides whether or not short groups of clicks are produced by one or more individuals from the following species: Blainville’s beaked whales, short- finned pilot whales, and Risso’s dolphins. The system locates individual clicks using the Teager energy operator and then constructs feature vectors for these clicks using cepstral analysis. Two different types of detectors confirm or reject the presence of each species. Gaussian mixture models (GMMs) are used to model time series independent characteristics of the species feature vector distributions. Support vector machines (SVMs) are used to model the boundaries between each species’ feature distribution and that of other species. Detection error tradeoff curves for all three species are shown with the following equal error rates: Blainville’s beaked whales (GMM 3.32%/SVM 5.54%), pilot whales (GMM 16.18%/SVM 15.00%), and Risso’s dolphins (GMM 0.03%/SVM 0.70%).

Roch, MA, Stinner-Sloan J, Baumann-Pickering S, Wiggins SM.  2015.  Compensating for the effects of site and equipment variation on delphinid species identification from their echolocation clicks. Journal of the Acoustical Society of America. 137:22-29.   10.1121/1.4904507   AbstractWebsite

A concern for applications of machine learning techniques to bioacoustics is whether or not classifiers learn the categories for which they were trained. Unfortunately, information such as characteristics of specific recording equipment or noise environments can also be learned. This question is examined in the context of identifying delphinid species by their echolocation clicks. To reduce the ambiguity between species classification performance and other confounding factors, species whose clicks can be readily distinguished were used in this study: Pacific white-sided and Risso's dolphins. A subset of data from autonomous acoustic recorders located at seven sites in the Southern California Bight collected between 2006 and 2012 was selected. Cepstral-based features were extracted for each echolocation click and Gaussian mixture models were used to classify groups of 100 clicks. One hundred Monte-Carlo three-fold experiments were conducted to examine classification performance where fold composition was determined by acoustic encounter, recorder characteristics, or recording site. The error rate increased from 6.1% when grouped by acoustic encounter to 18.1%, 46.2%, and 33.2% for grouping by equipment, equipment category, and site, respectively. A noise compensation technique reduced error for these grouping schemes to 2.7%, 4.4%, 6.7%, and 11.4%, respectively, a reduction in error rate of 56%-86%. (C) 2015 Acoustical Society of America.

Roth, EH, Hildebrand JA, Wiggins SM, Ross D.  2012.  Underwater ambient noise on the Chukchi Sea continental slope from 2006-2009. Journal of the Acoustical Society of America. 131:104-110.   10.1121/1.3664096   AbstractWebsite

From September 2006 to June 2009, an autonomous acoustic recorder measured ambient noise north of Barrow, Alaska on the continental slope at 235 m depth, between the Chukchi and Beaufort Seas. Mean monthly spectrum levels, selected to exclude impulsive events, show that months with open-water had the highest noise levels (80-83 dB re: 1 mu Pa-2/Hz at 20-50 Hz), months with ice coverage had lower spectral levels (70 dB at 50 Hz), and months with both ice cover and low wind speeds had the lowest noise levels (65 dB at 50 Hz). During ice covered periods in winter-spring there was significant transient energy between 10 and 100 Hz from ice fracture events. During ice covered periods in late spring there were significantly fewer transient events. Ambient noise increased with wind speed by similar to 1 dB/m/s for relatively open-water (0%-25% ice cover) and by similar to 0.5 dB/m/s for nearly complete ice cover (> 75%). In September and early October for all years, mean noise levels were elevated by 2-8 dB due to the presence of seismic surveys in the Chukchi and Beaufort Seas. (C) 2012 Acoustical Society of America. [DOI: 10.1121/1.3664096]

Roth, EH, Schmidt V, Hildebrand JA, Wiggins SM.  2013.  Underwater radiated noise levels of a research icebreaker in the central Arctic Ocean. Journal of the Acoustical Society of America. 133:1971-1980.   10.1121/1.4790356   AbstractWebsite

U.S. Coast Guard Cutter Healy's underwater radiated noise signature was characterized in the central Arctic Ocean during different types of ice-breaking operations. Propulsion modes included transit in variable ice cover, breaking heavy ice with backing-and-ramming maneuvers, and dynamic positioning with the bow thruster in operation. Compared to open-water transit, Healy's noise signature increased approximately 10 dB between 20 Hz and 2 kHz when breaking ice. The highest noise levels resulted while the ship was engaged in backing-and-ramming maneuvers, owing to cavitation when operating the propellers astern or in opposing directions. In frequency bands centered near 10, 50, and 100 Hz, source levels reached 190-200 dB re: 1 mu Pa at 1m (full octave band) during ice-breaking operations. (C) 2013 Acoustical Society of America.