LSTM ile EEG Tabanlı Kimliklendirme Sistemi Tasarımı

Tanımlama sistemleri son derece güvenilir kişisel veriler kullanılarak tasarlanmaktadır. Doğruluk oranı ve güvenilirlik bu sistemlerin en temel parametreleridir. Elektroensefalografi (EEG) sinyali zamana, içsel ve çevresel faktörlere bağlı olarak değişir. Yapılan çalışmalar sonucunda EEG sinyalinin tanımlama sistemlerinde kullanılabilirliği teyit edilmiştir. Çevresel etkiler en aza indirildiğinde vücut tarafından üretilen sinyallerin kişiselleştirilmiş sinyaller olduğu anlaşılmaktadır. Uzun Kısa Süreli Bellek (LSTM) yönteminin zaman serilerinde başarılı sonuçlar verdiği bilinmektedir. Bu çalışmada derin öğrenme tekniklerinden biri olan LSTM yöntemi kullanılarak bir tanımlama sistemi tasarlanmıştır. LSTM kullanılmadan once, EEG bazı işlemler ile frekans alt bileşenlerine bölünür. Bu ayrılan frekans alt bileşenlerinin korelasyon analizi ile delta dalgasının kullanılmasına karar verilmiştir. Hazırlanan system farklı koşullar altında incelenmiştir. Üç farklı eğitim serisi üzerinde 200 test yapılmıştır. En yüksek doğruluk oranı %89,5’tir. Ortalama doğruluk oranı %86,292’dir. Hazırlanan system farklı koşullar altında çalışacak şekilde tasarlanmıştır. Sistem çeşitli optimizasyon algoritmalrı kullanılarak gelişime açıktır.

Anahtar Kelimeler:

EEG, Kimliklendirme, Derin Öğrenme, LSTM

EEG Based Identification System Design via LSTM

Identification systems are designed using highly reliable personal data. Accuracy rate and reliability are the most basic parameters of these systems. Electroencephalography (EEG) signal varies depending on time, internal and environmental factors. As a result of the studies, the usability of the EEG signal in identification systems has been confirmed. It is understood that the signals produced by the body are personalized signals when the environmental effects are minimized. Successful results are known in the time series of the Long-Short Term Memory (LSTM) method. In this study, an identification system was designed by using LSTM method, which is one of the deep learning techniques. Before the LSTM is used, the EEG is subdivided into frequency subcomponents through some operations. It was decided to use the delta wave with correlation analysis of these separated frequency subcomponents. The prepared system was examined under different conditions. A total of 200 tests were performed on 3 different training series. The highest accuracy rate is 89.5%. The average accuracy rate is 86,292%. The prepared system is designed to operate under different conditions. The system is open to development using various optimization algorithms.

Keywords:

EEG, Identification, Deep Learning, LSTM,

PDF

___

Altahat, S., Che, G., Tran, D., Ma, W., 2015. Analysing the Robust EEG Channel Set for Person Authentication. International Conference on Neural Information Processing. Springer: 162–173.
Ashby, C., Bhatia, A., Tenore, F., Vogelstein, J., 2011. Low-cost electroencephalogram (EEG) based authentication. In Neural Engineering (NER). 2011 5th International IEEE/EMBS Conference. IEEE: 442–445.
Azevedo, F. A., Carvalho, L. R., Grinberg, L. T., Farfel, J. M., Ferretti, R. E., Leite, R. E., Herculano‐Houzel, S., 2009. Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled‐up primate brain. Journal of Comparative Neurology, 513(5): 532-541.
Bashar, K., Chiaki, I., Yoshida, H., 2016. Human identification from brain EEG signals using advanced machine learning method EEG-based biometrics. Biomedical Engineering and Sciences (IECBES), 2016 IEEE EMBS Conference. IEEE: 475–479.
Cireşan, D., Meier, U., Schmidhuber, J., 2012. Multi-column Deep Neural Networks for Image Classification. Computer Vision and Pattern Recognition: 3642-3649.
Deng. L., Yu, D., 2014. Deep Learning Methods and Applications. Foundations and Trends Signal Processing, (7): 198-250.
Florin, E., & Baillet, S. (2015). The brain's resting-state activity is shaped by synchronized cross-frequency coupling of neural oscillations. NeuroImage, 111, 26-35. doi:https://doi.org/10.1016/j.neuroimage.2015.01.054
Fukushima, K., 1980. Neocognitrion: a self organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biol. Cybern. 36 (4): 193-202.
Jasper, H., 1958. Report of the committee on methods of clinical examination in electroencephalography. Electroencephalography and Clinical Neurophysiology, 10: 370-375.
Kumari, P., Vaish, A., 2015. Brainwave based user identification system: A pilot study in robotics environment. Robotics and Autonomous Systems 65: 15–23.
Marcel, S., Millan, J., 2007. Person authentication using brainwaves (EEG) and maximum a posteriori model adaptation. IEEE transactions on pattern analysis and machine intelligence 29(4).
Mellinger, J., Schalk, G., Braun, C., Preissl, H., Rosenstiel, W., Birbaumer, N., Kübler, A., 2007. An MEG-based brain–computer interface (BCI). Neuroimage, 36(3): 581-593.
Pomas, K., Vinod, P., 2016. Utilizing individual alpha frequency and delta band power in EEG based biometric recognition. Systems, Man, and Cybernetics (SMC), 2016 IEEE International Conference. IEEE, 004787–004791
Schalk, G., Leuthardt, E.C., 2011. Brain-computer interfaces using electrocorticographic signals. IEEE Reviews in Biomedical Engineering, 4: 140-154.
Sohankar, J., Sadeghi, K., Banerjee, A., Gupta, S., 2015. E-bias: A pervasive eeg-based identification andauthentication system. Proceedings of the 11th ACM Symposium on QoS and Security for Wireless and Mobile Networks. ACM: 165–172.