A Combined CNN Architecture for Speech Emotion Recognition
Sensors · 2024
DOI: 10.3390/s24175797 ↗Kurz gefasst
Fachartikel in der Zeitschrift Sensors, 2024. Verfasst von Rolinson Begazo und 3 weiteren Autor:innen. Beteiligte Einrichtung: Electrical and Electronics Engineering Department, Universidad Católica San Pablo, Arequipa 04001, Peru. Hauptthemen laut MEDLINE: Emotions, Neural Networks, Computer, Deep Learning. Die Arbeit wurde bisher 28-mal zitiert (Crossref) und nennt 70 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: San Pablo Catholic University, Valencian International University.
Abstract (Originalsprache)
Emotion recognition through speech is a technique employed in various scenarios of Human–Computer Interaction (HCI). Existing approaches have achieved significant results; however, limitations persist, with the quantity and diversity of data being more notable when deep learning techniques are used. The lack of a standard in feature selection leads to continuous development and experimentation. Choosing and designing the appropriate network architecture constitutes another challenge. This study addresses the challenge of recognizing emotions in the human voice using deep learning techniques, proposing a comprehensive approach, and developing preprocessing and feature selection stages while constructing a dataset called EmoDSc as a result of combining several available databases. The synergy between spectral features and spectrogram images is investigated. Independently, the weighted accuracy obtained using only spectral features was 89%, while using only spectrogram images, the weighted accuracy reached 90%. These results, although surpassing previous research, highlight the strengths and limitations when operating in isolation. Based on this exploration, a neural network architecture composed of a CNN1D, a CNN2D, and an MLP that fuses spectral features and spectogram images is proposed. The model, supported by the unified dataset EmoDSc, demonstrates a remarkable accuracy of 96%.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Electrical and Electronics Engineering Department, Universidad Católica San Pablo, Arequipa 04001, Peru.
- Escuela de Ingeniería Informática, Facultad de Ingeniería, Universidad de Valparaíso, Valparaíso 2340000, Chile.
- Interdisciplinary Center for Biomedical Research and Health Engineering "MEDING", Universidad de Valparaíso, Valparaíso 2340000, Chile.
- ESTIA Institute of Technology, University Bordeaux, 64210 Bidart, France.
- Grupo de Investigación en Ciencia de Datos, Universidad Internacional de Valencia, 46002 Valencia, Spain.
ORCID-Profile: Rolinson Begazo, Ana Aguilera, Irvin Dongo, Yudith Cardinale
Förderung
- San Pablo Catholic University
- Valencian International University
Zitieren
APA
Begazo, R., Aguilera, A., Dongo, I., & Cardinale, Y. (2024). A Combined CNN Architecture for Speech Emotion Recognition. Sensors. https://doi.org/10.3390/s24175797
Vancouver
Begazo R, Aguilera A, Dongo I, Cardinale Y. A Combined CNN Architecture for Speech Emotion Recognition. Sensors. 2024. doi:10.3390/s24175797
BibTeX
@article{Begazo2024,
author = {Begazo, Rolinson and Aguilera, Ana and Dongo, Irvin and Cardinale, Yudith},
title = {A Combined CNN Architecture for Speech Emotion Recognition},
journal = {Sensors},
year = {2024},
doi = {10.3390/s24175797}
}
EUVIMED