Reconhecimento de padrões biométricos utilizando máquinas de aprendizado profundo: elaboração de modelo de máquina de aprendizado sequencial

Abstract

The brain-computer interface is one of the emerging fields of human-computer interaction due to its broad spectrum of applications, especially those dealing with human cognition. In this assignment, electroencephalography (EEG) signals are used as a base data to classify the state of the eyes (open or closed) by applying Long Short Term Memory (LSTM) networks and variants. For benchmarking purposes, the eye-state EEG data set available at the University of California Irvine (UCI) Machine Learning repository was used. The results indicated that the model is applicable to the classification of the data and that its performance is reasonable compared to the more expensive models computationally. In addition, a data capture system prototype was developed based on the YouMake platform, accessible hardware developed at the institution to have easy access to biomedical data considering the possibility of obtaining data with both conventional (but with prohibitively high costs) and device developed (with lower and limited cost).