Fusing Fixed and Adaptive Multi-resolution Features: A DWT-EWT Approach for Improved Speech Emotion Classification

DOI

10.2991/978-94-6239-707-1_18How to use a DOI?

Keywords

DWT; EWT; SEC; Feature Fusion; Multi-resolution Analysis

Abstract

Speech emotion classification (SEC) is the automatic identification process of the emotional states that are inherent parts of any utterance with the help of computer programming with high potential applications in the domain of medicine, security, surveillance, digital marketing, E-learning, internet search, personal communication, customer relation mechanisms, human-computer interaction, etc. Recent advances in speech emotion classification performance (ECP) employed various acoustic as well as non-acoustic features with the help of machine learning as well as deep learning algorithms. This paper introduces a new computing mechanism with the help of the hybrid approach of Discrete Wavelet Transform (DWT) and Empirical Wavelet Transform (EWT) with the intention to increase the classification accuracy level with the help of the proposed hybrid signal decomposition and feature fusion technique. Speech signals are broken into frames, that are then decomposed into four modes with the help of the proposed approach. i.e using DWT and EWT, followed by the extraction of five different entropy-based features namely “Approximate Entropy (ApE)”, “Permutation Entropy (PrE)”, “Increment Entropy (InE)”, “Sample Entropy (SaE)”, “Spectral Entropy (SpE)”, collectively termed Hybrid-Entropy (HEn) features and HMFCC features (Hybrid-Mel-Frequency Cepstral Coefficient). Experimental evaluation using a deep neural network (DNN) classifier demonstrates that combining HEn with HMFCC features derived from both decomposed modes achieves superior performance, attaining an accuracy of 89.76% on the EMODB dataset.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Devi Prasad Pattnaik
AU  - Bala Sai Srilatha Indira Dutt Vemuri
PY  - 2026
DA  - 2026/06/18
TI  - Fusing Fixed and Adaptive Multi-resolution Features: A DWT-EWT Approach for Improved Speech Emotion Classification
BT  - Proceedings of the International Conference on Recent Advances in Intelligent and Sustainable Technologies (RAIST 2026)
PB  - Atlantis Press
SP  - 211
EP  - 221
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6239-707-1_18
DO  - 10.2991/978-94-6239-707-1_18
ID  - Pattnaik2026
ER  -