Research on Emotion Recognition based on EEG Signal
Advisor: Prof. Jing Wang, March 2019 - June 2020
- Proposed a two-stream 3D Dense network, which fuses the spatial-spectral-temporal information of EEG signals in a unified network framework based on the constructed 3D EEG representation.
- Developed a parallel Spatial-Spectral/Temporal attention mechanism to adaptively capture discriminative patterns in brain regions, frequency bands and time stamps.
- Conducted extensive experiments on two benchmark datasets and the experimental results show that our SST-EmotionNet outperforms the state-of-the-art model by 1.78% on accuracy.
- Rated as a national-level project in the 2019 University Student Innovation and Entrepreneurship Training Program, Beijing Jiaotong University.
- Published a paper on 28th ACM International Conference on Multimedia.
