Speech emotion recognition using gammatone cepstral coefficients and deep learning features

Roneel V. Sharan

doi:10.1109/ICMLANT59547.2023.10372986

Speech emotion recognition using gammatone cepstral coefficients and deep learning features

Roneel V. Sharan^*

^*Corresponding author for this work

Australian Institute of Health Innovation

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

2 Citations (Scopus)

Abstract

Speech emotion recognition finds various applications, such as enhancing human-computer interaction and aiding remote mental health monitoring. This work proposes a method for speech emotion recognition using a combination of handcrafted and deep learning features. In particular, it studies the use of gammatone cepstral coefficients, which make use of gammatone filters which model the human auditory filters, and deep learning feature embeddings extracted from a pretrained network for audio analysis. A multilayer perceptron is employed for classification on the combined feature set where feature selection is performed using one-way analysis of variance. The proposed method is evaluated on a dataset of 535 speech recordings containing 7 types of emotions from 10 subjects. An average accuracy of 0.7631 is achieved in classifying the emotions using speech in leave-one-subject-out cross-validation. Analysis of the results shows that the use of gammatone cepstral coefficients provides improvement in classification accuracy over the conventional mel-frequency cepstral coefficients and the accuracy improves when combined with deep learning features.

Original language	English
Title of host publication	Proceedings of the 2023 IEEE International Conference on Machine Learning and Applied Network Technologies
Subtitle of host publication	ICMLANT 2023
Editors	Manuel Cardona, Vijender K. Solanki
Place of Publication	El Salvador
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Number of pages	4
ISBN (Electronic)	9798350303919
ISBN (Print)	9798350303926
DOIs	https://doi.org/10.1109/ICMLANT59547.2023.10372986
Publication status	Published - 2023
Event	2023 IEEE International Conference on Machine Learning and Applied Network Technologies, ICMLANT 2023 - Virtual, Online, El Salvador Duration: 14 Dec 2023 → 15 Dec 2023

Publication series

Name	IEEE International Conference on Machine Learning and Applied Network Technologies
Publisher	IEEE

Conference

Conference	2023 IEEE International Conference on Machine Learning and Applied Network Technologies, ICMLANT 2023
Country/Territory	El Salvador
City	Virtual, Online
Period	14/12/23 → 15/12/23

Keywords

deep learning features
feature selection
gammatone cepstral coefficients
mel-spectrogram
multilayer perceptron

Access to Document

10.1109/ICMLANT59547.2023.10372986

Cite this

Sharan, R. V. (2023). Speech emotion recognition using gammatone cepstral coefficients and deep learning features. In M. Cardona, & V. K. Solanki (Eds.), Proceedings of the 2023 IEEE International Conference on Machine Learning and Applied Network Technologies: ICMLANT 2023 ( IEEE International Conference on Machine Learning and Applied Network Technologies). Institute of Electrical and Electronics Engineers (IEEE). https://doi.org/10.1109/ICMLANT59547.2023.10372986

Sharan, Roneel V. / Speech emotion recognition using gammatone cepstral coefficients and deep learning features. Proceedings of the 2023 IEEE International Conference on Machine Learning and Applied Network Technologies: ICMLANT 2023. editor / Manuel Cardona ; Vijender K. Solanki. El Salvador : Institute of Electrical and Electronics Engineers (IEEE), 2023. ( IEEE International Conference on Machine Learning and Applied Network Technologies).

@inproceedings{a87a4b6844364395bff520dfc7595912,

title = "Speech emotion recognition using gammatone cepstral coefficients and deep learning features",

abstract = "Speech emotion recognition finds various applications, such as enhancing human-computer interaction and aiding remote mental health monitoring. This work proposes a method for speech emotion recognition using a combination of handcrafted and deep learning features. In particular, it studies the use of gammatone cepstral coefficients, which make use of gammatone filters which model the human auditory filters, and deep learning feature embeddings extracted from a pretrained network for audio analysis. A multilayer perceptron is employed for classification on the combined feature set where feature selection is performed using one-way analysis of variance. The proposed method is evaluated on a dataset of 535 speech recordings containing 7 types of emotions from 10 subjects. An average accuracy of 0.7631 is achieved in classifying the emotions using speech in leave-one-subject-out cross-validation. Analysis of the results shows that the use of gammatone cepstral coefficients provides improvement in classification accuracy over the conventional mel-frequency cepstral coefficients and the accuracy improves when combined with deep learning features.",

keywords = "deep learning features, feature selection, gammatone cepstral coefficients, mel-spectrogram, multilayer perceptron",

author = "Sharan, {Roneel V.}",

year = "2023",

doi = "10.1109/ICMLANT59547.2023.10372986",

language = "English",

isbn = "9798350303926",

series = " IEEE International Conference on Machine Learning and Applied Network Technologies",

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

editor = "Manuel Cardona and Solanki, {Vijender K.}",

booktitle = "Proceedings of the 2023 IEEE International Conference on Machine Learning and Applied Network Technologies",

address = "United States",

note = "2023 IEEE International Conference on Machine Learning and Applied Network Technologies, ICMLANT 2023 ; Conference date: 14-12-2023 Through 15-12-2023",

}

Sharan, RV 2023, Speech emotion recognition using gammatone cepstral coefficients and deep learning features. in M Cardona & VK Solanki (eds), Proceedings of the 2023 IEEE International Conference on Machine Learning and Applied Network Technologies: ICMLANT 2023. IEEE International Conference on Machine Learning and Applied Network Technologies, Institute of Electrical and Electronics Engineers (IEEE), El Salvador, 2023 IEEE International Conference on Machine Learning and Applied Network Technologies, ICMLANT 2023, Virtual, Online, El Salvador, 14/12/23. https://doi.org/10.1109/ICMLANT59547.2023.10372986

Speech emotion recognition using gammatone cepstral coefficients and deep learning features. / Sharan, Roneel V.
Proceedings of the 2023 IEEE International Conference on Machine Learning and Applied Network Technologies: ICMLANT 2023. ed. / Manuel Cardona; Vijender K. Solanki. El Salvador: Institute of Electrical and Electronics Engineers (IEEE), 2023. ( IEEE International Conference on Machine Learning and Applied Network Technologies).

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

TY - GEN

T1 - Speech emotion recognition using gammatone cepstral coefficients and deep learning features

AU - Sharan, Roneel V.

PY - 2023

Y1 - 2023

N2 - Speech emotion recognition finds various applications, such as enhancing human-computer interaction and aiding remote mental health monitoring. This work proposes a method for speech emotion recognition using a combination of handcrafted and deep learning features. In particular, it studies the use of gammatone cepstral coefficients, which make use of gammatone filters which model the human auditory filters, and deep learning feature embeddings extracted from a pretrained network for audio analysis. A multilayer perceptron is employed for classification on the combined feature set where feature selection is performed using one-way analysis of variance. The proposed method is evaluated on a dataset of 535 speech recordings containing 7 types of emotions from 10 subjects. An average accuracy of 0.7631 is achieved in classifying the emotions using speech in leave-one-subject-out cross-validation. Analysis of the results shows that the use of gammatone cepstral coefficients provides improvement in classification accuracy over the conventional mel-frequency cepstral coefficients and the accuracy improves when combined with deep learning features.

AB - Speech emotion recognition finds various applications, such as enhancing human-computer interaction and aiding remote mental health monitoring. This work proposes a method for speech emotion recognition using a combination of handcrafted and deep learning features. In particular, it studies the use of gammatone cepstral coefficients, which make use of gammatone filters which model the human auditory filters, and deep learning feature embeddings extracted from a pretrained network for audio analysis. A multilayer perceptron is employed for classification on the combined feature set where feature selection is performed using one-way analysis of variance. The proposed method is evaluated on a dataset of 535 speech recordings containing 7 types of emotions from 10 subjects. An average accuracy of 0.7631 is achieved in classifying the emotions using speech in leave-one-subject-out cross-validation. Analysis of the results shows that the use of gammatone cepstral coefficients provides improvement in classification accuracy over the conventional mel-frequency cepstral coefficients and the accuracy improves when combined with deep learning features.

KW - deep learning features

KW - feature selection

KW - gammatone cepstral coefficients

KW - mel-spectrogram

KW - multilayer perceptron

UR - http://www.scopus.com/inward/record.url?scp=85183468887&partnerID=8YFLogxK

U2 - 10.1109/ICMLANT59547.2023.10372986

DO - 10.1109/ICMLANT59547.2023.10372986

M3 - Conference proceeding contribution

AN - SCOPUS:85183468887

SN - 9798350303926

T3 - IEEE International Conference on Machine Learning and Applied Network Technologies

BT - Proceedings of the 2023 IEEE International Conference on Machine Learning and Applied Network Technologies

A2 - Cardona, Manuel

A2 - Solanki, Vijender K.

PB - Institute of Electrical and Electronics Engineers (IEEE)

CY - El Salvador

T2 - 2023 IEEE International Conference on Machine Learning and Applied Network Technologies, ICMLANT 2023

Y2 - 14 December 2023 through 15 December 2023

ER -

Sharan RV. Speech emotion recognition using gammatone cepstral coefficients and deep learning features. In Cardona M, Solanki VK, editors, Proceedings of the 2023 IEEE International Conference on Machine Learning and Applied Network Technologies: ICMLANT 2023. El Salvador: Institute of Electrical and Electronics Engineers (IEEE). 2023. ( IEEE International Conference on Machine Learning and Applied Network Technologies). doi: 10.1109/ICMLANT59547.2023.10372986

Speech emotion recognition using gammatone cepstral coefficients and deep learning features

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this