Decoding Digital Emotions: Advancing Online Learning with Speech-Emotion Recognition Systems

Sherif Welsen, Yiyang Liu

Research output: Chapter in Book/Conference proceedingBook Chapterpeer-review

Abstract

This chapter introduces a novel system tailored for emotion recognition in speech within online educational platforms. Developed using MATLAB, this system harnesses cutting-edge machine learning methodologies, employing datasets from the Berlin Database of Emotional Speech (EmoDB) and the Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS). Precise detection and categorization of emotional expressions in speech are made possible by the hybrid model it employs, which combines Long Short-Term Memory (LSTM) networks with one- and two-dimensional convolutional neural networks. This system effectively improves the interpretation of student emotions in virtual learning contexts, achieving an impressive accuracy rate of 83.95%. However, it is important to note that this work also underscores the necessity for ongoing research to further refine the system's performance and dependability. This endeavour marks a crucial advancement in customizing online education, aiming to foster more empathetic and engaging virtual learning environments.

Original languageEnglish
Title of host publicationLecture Notes in Educational Technology
PublisherSpringer Science and Business Media Deutschland GmbH
Pages139-151
Number of pages13
DOIs
Publication statusPublished - 2025

Publication series

NameLecture Notes in Educational Technology
VolumePart F312
ISSN (Print)2196-4963
ISSN (Electronic)2196-4971

Keywords

  • Convolution Neural Networks
  • Intelligent Tutoring
  • Long Short-Term Memory
  • Smart Campus
  • Speech Emotion Recognition (SER)

ASJC Scopus subject areas

  • Education
  • Computer Science Applications
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Decoding Digital Emotions: Advancing Online Learning with Speech-Emotion Recognition Systems'. Together they form a unique fingerprint.

Cite this