Voice pathology detection using auto-correlation of different filters bank

Ahmed Al-Nasheri, Zulfiqar Ali, Ghulam Muhammad, Mansour Alsulaiman

Research output: Contribution to conferencePaper

8 Citations (Scopus)

Abstract

This paper investigates the contribution of frequency bands for automatic voice pathology detection. First, the input voice signal is passed through a number of time-domain band-pass filters. The center frequencies are spaced on an octave scale. Each filter output is then divided into overlapping frames. Auto-correlation function is applied to each block to find the first largest peak, in areas other than near the dc value, and its corresponding lag. Therefore, each frame is having only these two features (peak value and lag). As classifier, we use Gaussian mixture models (GMM) and support vector machine (SVM), separately. Two well-known available databases, one in English (MEEI) and the other one in German (SVD), are used in the investigation. The results demonstrate that the most significant frequency range to detect voice pathology is between 1500 Hz and 3500 Hz. Using this filter band and with only two features, the accuracy is above 97% in case of the MEEI database.

Conference

Conference2014 11th IEEE/ACS International Conference on Computer Systems and Applications, AICCSA 2014
CountryQatar
CityDoha
Period10/11/1413/11/14

Fingerprint

Filter banks
Pathology
Autocorrelation
Singular value decomposition
Bandpass filters
Speech recognition
Frequency bands
Support vector machines
Classifiers

Keywords

  • Auto-correlation
  • GMM
  • SVD
  • SVM
  • voice pathology detection

Cite this

Al-Nasheri, A., Ali, Z., Muhammad, G., & Alsulaiman, M. (2015). Voice pathology detection using auto-correlation of different filters bank. 50-55. Paper presented at 2014 11th IEEE/ACS International Conference on Computer Systems and Applications, AICCSA 2014, Doha, Qatar. https://doi.org/10.1109/AICCSA.2014.7073178
Al-Nasheri, Ahmed ; Ali, Zulfiqar ; Muhammad, Ghulam ; Alsulaiman, Mansour. / Voice pathology detection using auto-correlation of different filters bank. Paper presented at 2014 11th IEEE/ACS International Conference on Computer Systems and Applications, AICCSA 2014, Doha, Qatar.6 p.
@conference{5c44b9d12ed04ed7a0a64d3c537d9d5a,
title = "Voice pathology detection using auto-correlation of different filters bank",
abstract = "This paper investigates the contribution of frequency bands for automatic voice pathology detection. First, the input voice signal is passed through a number of time-domain band-pass filters. The center frequencies are spaced on an octave scale. Each filter output is then divided into overlapping frames. Auto-correlation function is applied to each block to find the first largest peak, in areas other than near the dc value, and its corresponding lag. Therefore, each frame is having only these two features (peak value and lag). As classifier, we use Gaussian mixture models (GMM) and support vector machine (SVM), separately. Two well-known available databases, one in English (MEEI) and the other one in German (SVD), are used in the investigation. The results demonstrate that the most significant frequency range to detect voice pathology is between 1500 Hz and 3500 Hz. Using this filter band and with only two features, the accuracy is above 97{\%} in case of the MEEI database.",
keywords = "Auto-correlation, GMM, SVD, SVM, voice pathology detection",
author = "Ahmed Al-Nasheri and Zulfiqar Ali and Ghulam Muhammad and Mansour Alsulaiman",
year = "2015",
month = "4",
day = "2",
doi = "10.1109/AICCSA.2014.7073178",
language = "English",
pages = "50--55",
note = "2014 11th IEEE/ACS International Conference on Computer Systems and Applications, AICCSA 2014 ; Conference date: 10-11-2014 Through 13-11-2014",

}

Al-Nasheri, A, Ali, Z, Muhammad, G & Alsulaiman, M 2015, 'Voice pathology detection using auto-correlation of different filters bank' Paper presented at 2014 11th IEEE/ACS International Conference on Computer Systems and Applications, AICCSA 2014, Doha, Qatar, 10/11/14 - 13/11/14, pp. 50-55. https://doi.org/10.1109/AICCSA.2014.7073178

Voice pathology detection using auto-correlation of different filters bank. / Al-Nasheri, Ahmed; Ali, Zulfiqar; Muhammad, Ghulam; Alsulaiman, Mansour.

2015. 50-55 Paper presented at 2014 11th IEEE/ACS International Conference on Computer Systems and Applications, AICCSA 2014, Doha, Qatar.

Research output: Contribution to conferencePaper

TY - CONF

T1 - Voice pathology detection using auto-correlation of different filters bank

AU - Al-Nasheri, Ahmed

AU - Ali, Zulfiqar

AU - Muhammad, Ghulam

AU - Alsulaiman, Mansour

PY - 2015/4/2

Y1 - 2015/4/2

N2 - This paper investigates the contribution of frequency bands for automatic voice pathology detection. First, the input voice signal is passed through a number of time-domain band-pass filters. The center frequencies are spaced on an octave scale. Each filter output is then divided into overlapping frames. Auto-correlation function is applied to each block to find the first largest peak, in areas other than near the dc value, and its corresponding lag. Therefore, each frame is having only these two features (peak value and lag). As classifier, we use Gaussian mixture models (GMM) and support vector machine (SVM), separately. Two well-known available databases, one in English (MEEI) and the other one in German (SVD), are used in the investigation. The results demonstrate that the most significant frequency range to detect voice pathology is between 1500 Hz and 3500 Hz. Using this filter band and with only two features, the accuracy is above 97% in case of the MEEI database.

AB - This paper investigates the contribution of frequency bands for automatic voice pathology detection. First, the input voice signal is passed through a number of time-domain band-pass filters. The center frequencies are spaced on an octave scale. Each filter output is then divided into overlapping frames. Auto-correlation function is applied to each block to find the first largest peak, in areas other than near the dc value, and its corresponding lag. Therefore, each frame is having only these two features (peak value and lag). As classifier, we use Gaussian mixture models (GMM) and support vector machine (SVM), separately. Two well-known available databases, one in English (MEEI) and the other one in German (SVD), are used in the investigation. The results demonstrate that the most significant frequency range to detect voice pathology is between 1500 Hz and 3500 Hz. Using this filter band and with only two features, the accuracy is above 97% in case of the MEEI database.

KW - Auto-correlation

KW - GMM

KW - SVD

KW - SVM

KW - voice pathology detection

UR - http://www.scopus.com/inward/record.url?scp=84988239949&partnerID=8YFLogxK

U2 - 10.1109/AICCSA.2014.7073178

DO - 10.1109/AICCSA.2014.7073178

M3 - Paper

SP - 50

EP - 55

ER -

Al-Nasheri A, Ali Z, Muhammad G, Alsulaiman M. Voice pathology detection using auto-correlation of different filters bank. 2015. Paper presented at 2014 11th IEEE/ACS International Conference on Computer Systems and Applications, AICCSA 2014, Doha, Qatar. https://doi.org/10.1109/AICCSA.2014.7073178