Subband spectral histogram feature for improved sound recognition in low SNR conditions

Roneel V. Sharan; Tom J. Moir

doi:10.1109/ICDSP.2015.7251908

Subband spectral histogram feature for improved sound recognition in low SNR conditions

Roneel V. Sharan, Tom J. Moir

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding contribution › peer-review

3 Citations (Scopus)

Abstract

In this work, we use the subband intensity histogram values extracted from the spectrogram image of sound signals to form the feature vector for sound classification in an audio surveillance application. We propose two features based on this approach. Firstly, we extract the histogram features from the short time Fourier transform spectrogram image of sound signals, which we refer as the spectral histogram feature (SHF). Secondly, we apply the mel-filter to the spectrogram image before histogram feature extraction which we refer as the mel-spectral histogram feature (MSHF). When compared to baseline features from similar work, the SHF was shown to give significantly improved results in low SNR conditions with a higher overall classification performance. In addition, the MSHF produced even better results than the SHF with the added advantage of a lower feature dimension.

Original language	English
Title of host publication	2015 IEEE International Conference on Digital Signal Processing, DSP 2015
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Pages	432-435
Number of pages	4
Volume	2015-September
ISBN (Electronic)	9781479980581
DOIs	https://doi.org/10.1109/ICDSP.2015.7251908
Publication status	Published - 9 Sept 2015
Externally published	Yes
Event	IEEE International Conference on Digital Signal Processing, DSP 2015 - Singapore, Singapore Duration: 21 Jul 2015 → 24 Jul 2015

Conference

Conference	IEEE International Conference on Digital Signal Processing, DSP 2015
Country/Territory	Singapore
City	Singapore
Period	21/07/15 → 24/07/15

Keywords

audio surveillance
mel-filter
sound recognition
spectral histogram feature
support vector machine

Access to Document

10.1109/ICDSP.2015.7251908

Cite this

@inproceedings{86f13641225d4586a362cdf6142f623f,

title = "Subband spectral histogram feature for improved sound recognition in low SNR conditions",

abstract = "In this work, we use the subband intensity histogram values extracted from the spectrogram image of sound signals to form the feature vector for sound classification in an audio surveillance application. We propose two features based on this approach. Firstly, we extract the histogram features from the short time Fourier transform spectrogram image of sound signals, which we refer as the spectral histogram feature (SHF). Secondly, we apply the mel-filter to the spectrogram image before histogram feature extraction which we refer as the mel-spectral histogram feature (MSHF). When compared to baseline features from similar work, the SHF was shown to give significantly improved results in low SNR conditions with a higher overall classification performance. In addition, the MSHF produced even better results than the SHF with the added advantage of a lower feature dimension.",

keywords = "audio surveillance, mel-filter, sound recognition, spectral histogram feature, support vector machine",

author = "Sharan, {Roneel V.} and Moir, {Tom J.}",

year = "2015",

month = sep,

day = "9",

doi = "10.1109/ICDSP.2015.7251908",

language = "English",

volume = "2015-September",

pages = "432--435",

booktitle = "2015 IEEE International Conference on Digital Signal Processing, DSP 2015",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

address = "United States",

note = "IEEE International Conference on Digital Signal Processing, DSP 2015 ; Conference date: 21-07-2015 Through 24-07-2015",

}

Sharan, RV & Moir, TJ 2015, Subband spectral histogram feature for improved sound recognition in low SNR conditions. in 2015 IEEE International Conference on Digital Signal Processing, DSP 2015. vol. 2015-September, 7251908, Institute of Electrical and Electronics Engineers (IEEE), pp. 432-435, IEEE International Conference on Digital Signal Processing, DSP 2015, Singapore, Singapore, 21/07/15. https://doi.org/10.1109/ICDSP.2015.7251908

Subband spectral histogram feature for improved sound recognition in low SNR conditions. / Sharan, Roneel V.; Moir, Tom J.
2015 IEEE International Conference on Digital Signal Processing, DSP 2015. Vol. 2015-September Institute of Electrical and Electronics Engineers (IEEE), 2015. p. 432-435 7251908.

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding contribution › peer-review

TY - GEN

T1 - Subband spectral histogram feature for improved sound recognition in low SNR conditions

AU - Sharan, Roneel V.

AU - Moir, Tom J.

PY - 2015/9/9

Y1 - 2015/9/9

N2 - In this work, we use the subband intensity histogram values extracted from the spectrogram image of sound signals to form the feature vector for sound classification in an audio surveillance application. We propose two features based on this approach. Firstly, we extract the histogram features from the short time Fourier transform spectrogram image of sound signals, which we refer as the spectral histogram feature (SHF). Secondly, we apply the mel-filter to the spectrogram image before histogram feature extraction which we refer as the mel-spectral histogram feature (MSHF). When compared to baseline features from similar work, the SHF was shown to give significantly improved results in low SNR conditions with a higher overall classification performance. In addition, the MSHF produced even better results than the SHF with the added advantage of a lower feature dimension.

AB - In this work, we use the subband intensity histogram values extracted from the spectrogram image of sound signals to form the feature vector for sound classification in an audio surveillance application. We propose two features based on this approach. Firstly, we extract the histogram features from the short time Fourier transform spectrogram image of sound signals, which we refer as the spectral histogram feature (SHF). Secondly, we apply the mel-filter to the spectrogram image before histogram feature extraction which we refer as the mel-spectral histogram feature (MSHF). When compared to baseline features from similar work, the SHF was shown to give significantly improved results in low SNR conditions with a higher overall classification performance. In addition, the MSHF produced even better results than the SHF with the added advantage of a lower feature dimension.

KW - audio surveillance

KW - mel-filter

KW - sound recognition

KW - spectral histogram feature

KW - support vector machine

UR - http://www.scopus.com/inward/record.url?scp=84961316054&partnerID=8YFLogxK

U2 - 10.1109/ICDSP.2015.7251908

DO - 10.1109/ICDSP.2015.7251908

M3 - Conference proceeding contribution

AN - SCOPUS:84961316054

VL - 2015-September

SP - 432

EP - 435

BT - 2015 IEEE International Conference on Digital Signal Processing, DSP 2015

PB - Institute of Electrical and Electronics Engineers (IEEE)

T2 - IEEE International Conference on Digital Signal Processing, DSP 2015

Y2 - 21 July 2015 through 24 July 2015

ER -

Subband spectral histogram feature for improved sound recognition in low SNR conditions

Abstract

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this