| Title |
LSTM Based Behavior Classification Deep Learning Algorithm Using EEG |
| Authors |
박상욱(Sang-Uk Park) ; 한지훈(Ji-Hoon Han) ; 홍선기(Sun-Ki Hong) |
| DOI |
https://doi.org/10.5370/KIEE.2021.70.12.1924 |
| Keywords |
EEG Signals; LSTM; STFT; Deep Learning; PCA; Data Mining; GAN; Pearson Correlation |
| Abstract |
Precise measurement and analysis of EEG required expensive specialized equipment and experts. In this study, a solution to this problem is proposed using a deep learning algorithm. First, the EEG signal data are learned using a 2-CH single-electrode sensor in the frontal lobe and an LSTM deep learning model. A data mining algorithm is used to select high-quality data from the measured data. In this process, the data shortage caused by data removal is supplemented by multiplying data using IOU GAN. With the Pearson correlation analysis algorithm, the frequency domain where common features between data occur is identified. From this, the data capacity is reduced and the learning speed is improved. As a result, it was confirmed that behavior identification using EEG is possible with a small number of sensors and the proposed algorithm. |