Emotion Recognition Model of EEG Signals Based on Double Attention Mechanism
Author Information
Author(s): Ma Yahong, Huang Zhentao, Yang Yuyao, Zhang Shanwen, Dong Qi, Wang Rongrong, Hu Liangliang, Lucchiari Claudio
Primary Institution: Xi’an Key Laboratory of High Precision Industrial Intelligent Vision Measurement Technology, School of Electronic Information, Xijing University, Xi’an, China
Hypothesis
The study proposes a multi-channel automatic classification model for emotion EEG signals named DACB, which is based on dual attention mechanisms.
Conclusion
The DACB model achieves high accuracy in emotion classification tasks, demonstrating outstanding performance and generalization ability.
Supporting Evidence
- The DACB model achieved an accuracy of 99.96% in single-shot validation tests on the SEED-IV dataset.
- The accuracy of the DACB model in 10-fold cross-validation tests was 99.73% on the SEED-IV dataset.
- The DACB model outperformed other models in emotion classification tasks.
- The model effectively captures spatial features through a one-dimensional SE module.
- The model extracts temporal features using a bidirectional long short-term memory network.
- The study evaluated the model's performance across both discrete and multidimensional emotion space models.
Takeaway
This study created a new model that helps computers understand human emotions by looking at brain signals, and it works really well.
Methodology
The study used a multi-channel automatic classification model for emotion EEG signals, incorporating dual attention mechanisms, convolutional neural networks, and bidirectional long short-term memory networks.
Potential Biases
The study does not specify risks of bias.
Limitations
The study does not specify limitations.
Participant Demographics
The SEED-IV dataset included EEG recordings from 15 participants, while the DREAMER dataset included 23 healthy subjects (9 females and 14 males).
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website