Effect on speech emotion classification of a feature selection approach using a convolutional neural network

Speech emotion recognition (SER) is a challenging issue because it is not clear which features are effective for classification. Emotionally related features are always extracted from speech signals for emotional classification. Handcrafted features are mainly used for emotional identification from...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Ammar Amjad, Lal Khan, Hsien-Tsung Chang
Formato:	article
Lenguaje:	EN
Publicado:	PeerJ Inc. 2021
Materias:	Speech emotion recognition Feature extraction Feature selection Convolutional neural network Mel-spectrogram Data augmentation Electronic computers. Computer science QA75.5-76.95
Acceso en línea:	https://doaj.org/article/bbf3afa8d45b4e9eb37979e8193d2f28
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:bbf3afa8d45b4e9eb37979e8193d2f28
record_format	dspace
spelling	oai:doaj.org-article:bbf3afa8d45b4e9eb37979e8193d2f282021-11-05T15:05:33ZEffect on speech emotion classification of a feature selection approach using a convolutional neural network10.7717/peerj-cs.7662376-5992https://doaj.org/article/bbf3afa8d45b4e9eb37979e8193d2f282021-11-01T00:00:00Zhttps://peerj.com/articles/cs-766.pdfhttps://peerj.com/articles/cs-766/https://doaj.org/toc/2376-5992Speech emotion recognition (SER) is a challenging issue because it is not clear which features are effective for classification. Emotionally related features are always extracted from speech signals for emotional classification. Handcrafted features are mainly used for emotional identification from audio signals. However, these features are not sufficient to correctly identify the emotional state of the speaker. The advantages of a deep convolutional neural network (DCNN) are investigated in the proposed work. A pretrained framework is used to extract the features from speech emotion databases. In this work, we adopt the feature selection (FS) approach to find the discriminative and most important features for SER. Many algorithms are used for the emotion classification problem. We use the random forest (RF), decision tree (DT), support vector machine (SVM), multilayer perceptron classifier (MLP), and k-nearest neighbors (KNN) to classify seven emotions. All experiments are performed by utilizing four different publicly accessible databases. Our method obtains accuracies of 92.02%, 88.77%, 93.61%, and 77.23% for Emo-DB, SAVEE, RAVDESS, and IEMOCAP, respectively, for speaker-dependent (SD) recognition with the feature selection method. Furthermore, compared to current handcrafted feature-based SER methods, the proposed method shows the best results for speaker-independent SER. For EMO-DB, all classifiers attain an accuracy of more than 80% with or without the feature selection technique.Ammar AmjadLal KhanHsien-Tsung ChangPeerJ Inc.articleSpeech emotion recognitionFeature extractionFeature selectionConvolutional neural networkMel-spectrogramData augmentationElectronic computers. Computer scienceQA75.5-76.95ENPeerJ Computer Science, Vol 7, p e766 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Speech emotion recognition Feature extraction Feature selection Convolutional neural network Mel-spectrogram Data augmentation Electronic computers. Computer science QA75.5-76.95
spellingShingle	Speech emotion recognition Feature extraction Feature selection Convolutional neural network Mel-spectrogram Data augmentation Electronic computers. Computer science QA75.5-76.95 Ammar Amjad Lal Khan Hsien-Tsung Chang Effect on speech emotion classification of a feature selection approach using a convolutional neural network
description	Speech emotion recognition (SER) is a challenging issue because it is not clear which features are effective for classification. Emotionally related features are always extracted from speech signals for emotional classification. Handcrafted features are mainly used for emotional identification from audio signals. However, these features are not sufficient to correctly identify the emotional state of the speaker. The advantages of a deep convolutional neural network (DCNN) are investigated in the proposed work. A pretrained framework is used to extract the features from speech emotion databases. In this work, we adopt the feature selection (FS) approach to find the discriminative and most important features for SER. Many algorithms are used for the emotion classification problem. We use the random forest (RF), decision tree (DT), support vector machine (SVM), multilayer perceptron classifier (MLP), and k-nearest neighbors (KNN) to classify seven emotions. All experiments are performed by utilizing four different publicly accessible databases. Our method obtains accuracies of 92.02%, 88.77%, 93.61%, and 77.23% for Emo-DB, SAVEE, RAVDESS, and IEMOCAP, respectively, for speaker-dependent (SD) recognition with the feature selection method. Furthermore, compared to current handcrafted feature-based SER methods, the proposed method shows the best results for speaker-independent SER. For EMO-DB, all classifiers attain an accuracy of more than 80% with or without the feature selection technique.
format	article
author	Ammar Amjad Lal Khan Hsien-Tsung Chang
author_facet	Ammar Amjad Lal Khan Hsien-Tsung Chang
author_sort	Ammar Amjad
title	Effect on speech emotion classification of a feature selection approach using a convolutional neural network
title_short	Effect on speech emotion classification of a feature selection approach using a convolutional neural network
title_full	Effect on speech emotion classification of a feature selection approach using a convolutional neural network
title_fullStr	Effect on speech emotion classification of a feature selection approach using a convolutional neural network
title_full_unstemmed	Effect on speech emotion classification of a feature selection approach using a convolutional neural network
title_sort	effect on speech emotion classification of a feature selection approach using a convolutional neural network
publisher	PeerJ Inc.
publishDate	2021
url	https://doaj.org/article/bbf3afa8d45b4e9eb37979e8193d2f28
work_keys_str_mv	AT ammaramjad effectonspeechemotionclassificationofafeatureselectionapproachusingaconvolutionalneuralnetwork AT lalkhan effectonspeechemotionclassificationofafeatureselectionapproachusingaconvolutionalneuralnetwork AT hsientsungchang effectonspeechemotionclassificationofafeatureselectionapproachusingaconvolutionalneuralnetwork
_version_	1718444187581415424

Effect on speech emotion classification of a feature selection approach using a convolutional neural network

Ejemplares similares