Augmented Audio Data in Improving Speech Emotion Classification Tasks

Nusrat J Shoumy; Li-Minn Ang; D. M. Motiur Rahaman; Tanveer Zia; Kah Phooi Seng; Sabira Khatun

doi:10.1007/978-3-030-79463-7_30

Back

Augmented Audio Data in Improving Speech Emotion Classification Tasks

Conference paper

Peer reviewed

Augmented Audio Data in Improving Speech Emotion Classification Tasks

Nusrat J Shoumy, Li-Minn Ang, D. M. Motiur Rahaman, Tanveer Zia, Kah Phooi Seng and Sabira Khatun

Advances and Trends in Artificial Intelligence. From Theory to Practice, pp.360-365

International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, 34th (Kuala Lumpur, Malaysia, 26-Jul-2021 - 29-Jul-2021)

Lecture Notes in Computer Science, 12799, Springer International Publishing

2021

DOI: https://doi.org/10.1007/978-3-030-79463-7_30

Files and links (1)

url

https://doi.org/10.1007/978-3-030-79463-7_30View

Published Version

Abstract

Audio data

Data augmentation

Data classification

Emotion recognition

Neural Network

To achieve high performance and classification accuracy, classification of emotions from audio or speech signals requires large quantities of data. Big datasets, however, are not always readily accessible. A good solution to this issue is to increase the data and augment it to construct a larger dataset for the classifier’s training. This paper proposes a unimodal approach that focuses on two main concepts: (1) augmenting speech signals to generate additional data samples; and (2) constructing classification models to identify emotion expressed through speech. In addition, three classifiers (Convolutional Neural Network (CNN), Naïve Bayes (NB) and K-Nearest Neighbor (kNN)) were further tested in order to decide which of the classifiers had the best results. We used augmented audio data from a dataset (SAVEE) in the proposed method to conduct training (50%), and testing (50%) was executed using the original data. The best performance of approximately 83% was found to be a mixture of augmentation strategies using the CNN classifier. Our proposed augmentation approach together with appropriate classification model enhances the efficiency of voice emotion recognition.

Details

Title: Augmented Audio Data in Improving Speech Emotion Classification Tasks
Authors: Nusrat J Shoumy (Corresponding Author) - Charles Sturt University
Li-Minn Ang (Author) - University of the Sunshine Coast, Queensland, School of Science, Technology and Engineering
D. M. Motiur Rahaman (Author) - Charles Sturt University
Tanveer Zia (Author) - Charles Sturt University
Kah Phooi Seng (Author) - UNSW Australia
Sabira Khatun (Author) - Universiti Malaysia Pahang
Publication details: Advances and Trends in Artificial Intelligence. From Theory to Practice, pp.360-365
Conference details: International Conference on Industrial, Engineering and Other Applications of Applied Intelligent Systems, 34th (Kuala Lumpur, Malaysia, 26-Jul-2021 - 29-Jul-2021)
Series: Lecture Notes in Computer Science; 12799
Publisher: Springer International Publishing
DOI: 10.1007/978-3-030-79463-7_30; 10.1007/978-3-030-79463-7
ISSN: 1611-3349; 1611-3349
ISBN: 9783030794637
Organisation Unit: School of Science, Technology and Engineering; University of the Sunshine Coast, Queensland; Engage Research Lab
Language: English
Record Identifier: 99571605802621
Output Type: Conference paper

Metrics

23 Record Views