Deep Learning for Personal Identification Using EEG

CNN-GRU and CNN-LSTM has been used for personal identification from different affective states when train on the state-of-the-art affective dataset DEAP. Both of them reach up to 99.90–100% mean Correct Recognition Rate.

For more details, please refer to: https://ieeexplore.ieee.org/document/8745473

Environment

• CUDA toolkit 10.0 and CuDNN v7
• Python 3.6.8
• tensorflow-gpu (1.14)
• matplotlib (1.5.3)
• scikit-learn (0.21.3)
• numpy (1.14.5)
• pandas (0.18.1)

Code Description (To be updated)

Data located locally on Vistec server at /Documents/gong/DEAP (10.204.142.25:/Documents/gong/DEAP/) Required Files X_0.npy,X_1.npy,X_2.npy,X_3.npy,X_4.npy Y_4classes.npy

Citation

Following citation format can be used for BibTex:

@ARTICLE{8745473,
author={T. {Wilaiprasitporn} and A. {Ditthapron} and K. {Matchaparn} and T. {Tongbuasirilai} and N. {Banluesombatkul} and E. {Chuangsuwanich}},
journal={IEEE Transactions on Cognitive and Developmental Systems},
title={Affective EEG-Based Person Identification Using the Deep Learning Approach},
year={2019},
pages={1-1},
doi={10.1109/TCDS.2019.2924648},
}