EmotionMIL: An End-to-End Multiple Instance Learning Framework for Emotion Recognition from EEG Signals

The code will be released once the paper is accepted.

Overview

The EmotionMIL framework for emotion recognition from multi-channel EEG signals. (a) EEG signal segmentation and preprocessing. (b) Temporal mixer layer for capturing temporal dependencies within EEG segments. (c) Spatial mixer layer for capturing spatial dependencies between EEG channels. (d) EEGMixer for instance feature extraction. (e) Multiple instance pooling layer for aggregating instance features and predicting the overall emotion label. (f) Detailed Rettention-based MIL pooling layer.

Citation

If you use the code or results in your research, please consider citing our work at:

@article{yu2024emotionmil,
  title={EmotionMIL: An End-to-End Multiple Instance Learning Framework for Emotion Recognition from EEG Signals},
  author={},
  journal={},
  year={2024},
  doi={},
  url={},
}