Scholarlink

This paper presents Emotion-H Net, a deep learning model designed to interpret emotional states from EEG signals. By combining transformer encoder modules with positional encodings, the architecture effectively captures the temporal patterns across different EEG channels. Standardisation and oversampling ensured balanced and clean inputs, while training employed the AdamW optimizer with a learning rate of 0.001 over 10 epochs. Evaluated on the DEAP dataset, the model achieved an accuracy of 86.82%, an F1-score of 0.86, and an AUROC of 0.977, demonstrating significant gains over conventional methods and strong potential for robust emotion recognition.

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

Emotion-H Net: A Hybrid Transformer Architecture for Emotion Recognition via Multi-feature Signal Fusion

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Aryan Goel,
  • Ankit Bhardwaj,
  • Sarthak Virmani,
  • Rohan Sharma,
  • Jyoti Yadav,
  • Sangeeta Gupta

摘要

This paper presents Emotion-H Net, a deep learning model designed to interpret emotional states from EEG signals. By combining transformer encoder modules with positional encodings, the architecture effectively captures the temporal patterns across different EEG channels. Standardisation and oversampling ensured balanced and clean inputs, while training employed the AdamW optimizer with a learning rate of 0.001 over 10 epochs. Evaluated on the DEAP dataset, the model achieved an accuracy of 86.82%, an F1-score of 0.86, and an AUROC of 0.977, demonstrating significant gains over conventional methods and strong potential for robust emotion recognition.