Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Acoustic detection and classification of Microchiroptera using machine learning: lessons learned from automatic

Mark D Skowronski1, John G Harris

  • 1Computational Neuro-Engineering Lab, Electrical and Computer Engineering, University of Florida, Gainesville, Florida 32611, USA. markskow@cnel.ufl.edu

The Journal of the Acoustical Society of America
|April 6, 2006
PubMed
Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Editorial: Bio-inspired Audio Processing, Models and Systems.

Frontiers in neuroscience·2019
Same author

Perceptual and Quantitative Assessment of Dysphonia Across Vowel Categories.

Journal of voice : official journal of the Voice Foundation·2018
Same author

Comparison of Pitch Strength With Perceptual and Other Acoustic Metric Outcome Measures Following Medialization Laryngoplasty.

Journal of voice : official journal of the Voice Foundation·2018
Same author

The Perception of Breathiness in the Voices of Pediatric Speakers.

Journal of voice : official journal of the Voice Foundation·2017
Same author

Pitch Strength as an Outcome Measure for Treatment of Dysphonia.

Journal of voice : official journal of the Voice Foundation·2017
Same author

Cepstral Peak Sensitivity: A Theoretic Analysis and Comparison of Several Implementations.

Journal of voice : official journal of the Voice Foundation·2015

Machine learning significantly improves bat call detection and classification accuracy. This approach offers an order of magnitude reduction in errors compared to traditional methods for microchiroptera acoustic analysis.

Area of Science:

  • Bioacoustics
  • Machine Learning
  • Computational Biology

Background:

  • Current microchiroptera acoustic analysis relies on global call features, similar to early human speech recognition.
  • This expert-driven approach struggles with signal variation, limiting accuracy.

Purpose of the Study:

  • To compare machine learning algorithms against conventional methods for automatic acoustic detection and classification of microchiroptera.
  • To evaluate the efficacy of Gaussian mixture models (GMM) and hidden Markov models (HMM) in bat call analysis.

Main Methods:

  • Hand-labeled search-phase calls from five bat species were used to train GMM and HMM models.
  • Performance was assessed by comparing error rates of machine learning models against baseline detectors and classifiers.

Related Experiment Videos

Main Results:

  • The GMM detector achieved 4% error, significantly outperforming a broadband energy detector (32% error).
  • GMM and HMM classifiers yielded very low error rates (0.6% and 16.9% respectively) compared to discriminant function analysis (16.9%).
  • Machine learning methods demonstrated an order of magnitude lower error rate than conventional techniques.

Conclusions:

  • Machine learning algorithms offer a substantial improvement in the accuracy of automatic acoustic detection and classification of microchiroptera.
  • This study highlights the potential of machine learning to overcome limitations of expert-knowledge-based approaches in bioacoustics.