Acoustic Event Classification using Low-Resolution Multi-label Non-negative Matrix Deconvolution
×
Cite This
Citation & Abstract
L. Vuegen, P. Karsmakers, B. Vanrumste, and HU. VA. Hamme, "Acoustic Event Classification using Low-Resolution Multi-label Non-negative Matrix Deconvolution," J. Audio Eng. Soc., vol. 66, no. 5, pp. 369-384, (2018 May.). doi: https://doi.org/10.17743/jaes.2018.0018
L. Vuegen, P. Karsmakers, B. Vanrumste, and HU. VA. Hamme, "Acoustic Event Classification using Low-Resolution Multi-label Non-negative Matrix Deconvolution," J. Audio Eng. Soc., vol. 66 Issue 5 pp. 369-384, (2018 May.). doi: https://doi.org/10.17743/jaes.2018.0018
Abstract: With the increased proliferation of interconnected devices that have built-in microphones, acoustic event classification and monitoring becomes possible in a wide variety of applications, such as surveillance, healthcare, military, machine diagnostics, and wildlife tracking. The promise and success of these applications depends on robust sensing of acoustic events in the environment. Typically, sound event classes are defined by annotating training data, which is a laborious process. This work introduces an extended version of non-negative matrix deconvolution (NMD), called low-resolution multi-label non-negative matrix deconvolution (LRM-NMD), where both the observation data and the available labeling information are used during training. The proposed extension of NMD was successfully applied to the classification of acoustic events even in noisy conditions with overlapping events. Low-resolution, multi-labeling information simply indicates that the sound classes of the events take place over a longer period of time in the acoustic data without identifying beginning or endings of the individual events.
@article{vuegen2018acoustic,
author={vuegen, lode and karsmakers, peter and vanrumste, bart and hamme, hugo van},
journal={journal of the audio engineering society},
title={acoustic event classification using low-resolution multi-label non-negative matrix deconvolution},
year={2018},
volume={66},
number={5},
pages={369-384},
doi={https://doi.org/10.17743/jaes.2018.0018},
month={may},}
@article{vuegen2018acoustic,
author={vuegen, lode and karsmakers, peter and vanrumste, bart and hamme, hugo van},
journal={journal of the audio engineering society},
title={acoustic event classification using low-resolution multi-label non-negative matrix deconvolution},
year={2018},
volume={66},
number={5},
pages={369-384},
doi={https://doi.org/10.17743/jaes.2018.0018},
month={may},
abstract={with the increased proliferation of interconnected devices that have built-in microphones, acoustic event classification and monitoring becomes possible in a wide variety of applications, such as surveillance, healthcare, military, machine diagnostics, and wildlife tracking. the promise and success of these applications depends on robust sensing of acoustic events in the environment. typically, sound event classes are defined by annotating training data, which is a laborious process. this work introduces an extended version of non-negative matrix deconvolution (nmd), called low-resolution multi-label non-negative matrix deconvolution (lrm-nmd), where both the observation data and the available labeling information are used during training. the proposed extension of nmd was successfully applied to the classification of acoustic events even in noisy conditions with overlapping events. low-resolution, multi-labeling information simply indicates that the sound classes of the events take place over a longer period of time in the acoustic data without identifying beginning or endings of the individual events.},}
TY - paper
TI - Acoustic Event Classification using Low-Resolution Multi-label Non-negative Matrix Deconvolution
SP - 369
EP - 384
AU - Vuegen, Lode
AU - Karsmakers, Peter
AU - Vanrumste, Bart
AU - Hamme, Hugo Van
PY - 2018
JO - Journal of the Audio Engineering Society
IS - 5
VO - 66
VL - 66
Y1 - May 2018
TY - paper
TI - Acoustic Event Classification using Low-Resolution Multi-label Non-negative Matrix Deconvolution
SP - 369
EP - 384
AU - Vuegen, Lode
AU - Karsmakers, Peter
AU - Vanrumste, Bart
AU - Hamme, Hugo Van
PY - 2018
JO - Journal of the Audio Engineering Society
IS - 5
VO - 66
VL - 66
Y1 - May 2018
AB - With the increased proliferation of interconnected devices that have built-in microphones, acoustic event classification and monitoring becomes possible in a wide variety of applications, such as surveillance, healthcare, military, machine diagnostics, and wildlife tracking. The promise and success of these applications depends on robust sensing of acoustic events in the environment. Typically, sound event classes are defined by annotating training data, which is a laborious process. This work introduces an extended version of non-negative matrix deconvolution (NMD), called low-resolution multi-label non-negative matrix deconvolution (LRM-NMD), where both the observation data and the available labeling information are used during training. The proposed extension of NMD was successfully applied to the classification of acoustic events even in noisy conditions with overlapping events. Low-resolution, multi-labeling information simply indicates that the sound classes of the events take place over a longer period of time in the acoustic data without identifying beginning or endings of the individual events.
With the increased proliferation of interconnected devices that have built-in microphones, acoustic event classification and monitoring becomes possible in a wide variety of applications, such as surveillance, healthcare, military, machine diagnostics, and wildlife tracking. The promise and success of these applications depends on robust sensing of acoustic events in the environment. Typically, sound event classes are defined by annotating training data, which is a laborious process. This work introduces an extended version of non-negative matrix deconvolution (NMD), called low-resolution multi-label non-negative matrix deconvolution (LRM-NMD), where both the observation data and the available labeling information are used during training. The proposed extension of NMD was successfully applied to the classification of acoustic events even in noisy conditions with overlapping events. Low-resolution, multi-labeling information simply indicates that the sound classes of the events take place over a longer period of time in the acoustic data without identifying beginning or endings of the individual events.
Authors:
Vuegen, Lode; Karsmakers, Peter; Vanrumste, Bart; Hamme, Hugo Van
Affiliations:
KU Leuven, Dpt. of Electrical Engineering, Leuven, Belgium; IMEC, Leuven, Belgium(See document for exact affiliation information.) JAES Volume 66 Issue 5 pp. 369-384; May 2018
Publication Date:
May 24, 2018Import into BibTeX
Permalink:
http://www.aes.org/e-lib/browse.cfm?elib=19567