Scholarlink

Interpreting speech emotions and understanding emotion states has significantly increased interest in research to improve the quality of different services and advance human-computer interaction. Recurrent-based neural networks can be utilized to extract the spatial and temporal emotion patterns of speech. In this paper, we empirically evaluate different recurrent-based neural network capabilities for efficient and precise recognition of different emotion patterns in speech. Our empirical analysis and study focus on determining the optimal recurrent architecture for speech emotion recognition for real-time applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Empirical Evaluation of Deep Recurrent Deep Neural Networks Models for Speech Emotion Recognition for Real Time Applications

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Nelly Elsayed,
  • Constantinos L. Zekios,
  • Zag ElSayed,
  • Ernest V. Pedapati,
  • Stavros Georgakopoulos

摘要

Interpreting speech emotions and understanding emotion states has significantly increased interest in research to improve the quality of different services and advance human-computer interaction. Recurrent-based neural networks can be utilized to extract the spatial and temporal emotion patterns of speech. In this paper, we empirically evaluate different recurrent-based neural network capabilities for efficient and precise recognition of different emotion patterns in speech. Our empirical analysis and study focus on determining the optimal recurrent architecture for speech emotion recognition for real-time applications.