Enhancing children’s emotional education with speech emotion recognition using deep learning
PeerJ Computer Science · 08.07.2026 · 12 · e3758
DOI: 10.7717/peerj-cs.3758 ↗Kurz gefasst
Fachartikel in der Zeitschrift PeerJ Computer Science, online erschienen am 08.07.2026. Verfasst von Ning Cui. Die Arbeit nennt 60 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Jiangsu Higher Education Institutions, Jiangsu Province’s Colleges and Universities.
Themen: Kinderkrankenpflege
Abstract (Originalsprache)
Speech emotion recognition (SER) plays a pivotal role in advancing affective computing and intelligent human–computer interaction, particularly in the context of children’s emotional communication. Traditional SER systems relying on handcrafted acoustic features often struggle to capture the complex temporal and spectral patterns inherent in emotional speech. To overcome these limitations, this study proposes a hybrid convolution-transformer framework based on the Vision Transformer (ViT) architecture for recognizing emotions from children’s speech. The proposed method first converts audio signals into mel-spectrograms and then utilizes ViT’s global self-attention mechanism to model both local spectral cues and long-range temporal dependencies. To ensure comprehensive validation, we conducted multi-level benchmarking experiments against classical machine learning models, deep learning baselines, and advanced architecture for speech emotion recognition. Experimental results demonstrate that the proposed ViT-based approach achieves competitive performance while maintaining a lightweight architecture. This balance between computational efficiency and modeling power highlights the potential of ViT for real-world SER applications.
Förderung
- Jiangsu Higher Education Institutions
- Jiangsu Province’s Colleges and Universities
Zitieren
APA
Cui, N. (2026). Enhancing children’s emotional education with speech emotion recognition using deep learning. PeerJ Computer Science, 12, e3758. https://doi.org/10.7717/peerj-cs.3758
Vancouver
Cui N. Enhancing children’s emotional education with speech emotion recognition using deep learning. PeerJ Computer Science. 2026;12:e3758. doi:10.7717/peerj-cs.3758
BibTeX
@article{Cui2026,
author = {Cui, Ning},
title = {Enhancing children’s emotional education with speech emotion recognition using deep learning},
journal = {PeerJ Computer Science},
year = {2026},
volume = {12},
pages = {e3758},
doi = {10.7717/peerj-cs.3758}
}
EUVIMED