EEG-based user identification system using 1D-convolutional long short-term memory neural networks

2020

Sci Rep

We present a four-objective optimization method for optimal electroencephalographic (eeG) channel selection to provide access to subjects with permission in a system by detecting intruders and identifying the subject. Each instance was represented by four features computed from two subbands, extracted using empirical mode decomposition (eMD) for each channel, and the feature vectors were used as input for one-class/multi-class support vector machines (SVMs). We tested the method on data from the event-related potentials (ERPs) of 26 subjects and 56 channels. The optimization process was performed by the non-dominated sorting genetic algorithm (NSGA), which found a threechannel combination that achieved an accuracy of 0.83, with both a true acceptance rate (TAR) and a true rejection rate (TRR) of 1.00. In the best case, we obtained an accuracy of up to 0.98 for subject identification with a TAR of 0.95 and a TRR 0.93, all using seven EEG channels found by NSGA-III in a subset of subjects manually created. The findings were also validated using 10 different subdivisions of subjects randomly created, obtaining up to 0.97 ± 0.02 of accuracy, a TAR of 0.81 ± 0.12 and TRR of 0.85 ± 0.10 using eight channels found by NSGA-III. These results support further studies on larger datasets for potential applications of EEG in identification and authentication systems.

mentioning

confidence: 99%

Multi-objective optimization for EEG channel selection and accurate intruder detection in an EEG-based subject identification system

2020

Sci Rep

“…Authors in [10] presented and approach using 1D convolutional long short-term memory neural network (1D-Convolutional LSTM) validating the method with a dataset of 119 subjects and 64 channels. They report an accuracy of 0.92 using only 4 channels, the validation was made using only 3-folds cross-validation.…”

Section: Related Workmentioning

confidence: 99%

Event-related potential from EEG for a two-step Identity Authentication System

2019 IEEE 17th International Conference on Industrial Informatics (INDIN)

2019

Current problems related to high-level security access are increasing, leaving organizations and persons unsafe. A recent good candidate to create a robust identity authentication system is based on brain signals recorded with electroencephalograms (EEG). In this paper, EEG-based brain signals of 56 channels, from event-related potentials (ERPs), are used for Subject identification. The ERPs are from positive or negative feedback-related responses of a P300-speller system. The feature extraction part was done with empirical mode decomposition (EMD) extracting 2 intrinsic mode functions (IMFs) per channel, that were selected based on the Minkowski distance. After that, 4 features are computed per IMF; 2 energy features (instantaneous and teager energy) and 2 fractal features (Higuchi and Petrosian fractal dimension). Support vector machine (SVM) was used for the classification stage with an accuracy index computed using 10folds cross-validation for evaluating the classifier's performance. Since high-density EEG information was available, the wellknown backward-elimination and forward-addition greedy algorithms were used to reduce or increase the number of channels, step by step. Using the proposed method for subject identification from a positive or negative feedback-related response and then identify the subject will add a layer to improve the security system. The results obtained show that subject identification is feasible even using a low number of channels: E.g., 0.89 of accuracy using 5 channels with a mixed population and 0.93 with a male-only population.

“…In general, there are a number of approaches that use raw data as input for various configurations of neural networks (NN) [13][14][15][16] and several have been proposed to tackle the high dimensionality of the data using methods for feature extraction, such as principal component analysis (PCA) 17 , the discrete wavelet transform (DWT) 3 , or empirical mode decomposition (EMD) 2,3,11,12 .…”

mentioning

confidence: 99%

“…Several approaches have been proposed for the creation of a biometric system following various experiment configurations, with various paradigms and methods for feature extraction and classification using the public EEG Motor Movement/Imagery Dataset (EEGMMIDB), using various configurations of neural networks 14,[18][19][20] , other supervised and unsupervised techniques 5,[21][22][23][24][25][26][27][28][29][30][31] , and methods for EEG channel selection 6,32,33 .…”

mentioning

confidence: 99%

“…To overcome the need for a large amount of data of deep learning approaches, there is an approach that uses simple data augmentation techniques by creating overlapped time windows 18 . Other related proposals using neural networks have been presented and compared in the state of the art [28][29][30][31] , where some of the most relevant works used around 100 subjects and in most of the cases 64 channels for testing their approaches 13,14,18,36 . However, there is no defined method for channel selection, since the process for selecting the most relevant channels requires the repetition of the classification process several times, and it is well known that deep learning approaches are computationally costly 30,34 .…”

mentioning

confidence: 99%

See 1 more Smart Citation

Towards a minimal EEG channel array for a biometric system using resting-state and a genetic algorithm for channel selection

2020

Sci Rep

We present a new approach for a biometric system based on electroencephalographic (EEG) signals of resting-state, that can identify a subject and reject intruders with a minimal subset of EEG channels. To select features, we first use the discrete wavelet transform (DWT) or empirical mode decomposition (EMD) to decompose the EEG signals into a set of sub-bands, for which we compute the instantaneous and Teager energy and the Higuchi and Petrosian fractal dimensions for each sub-band. The obtained features are used as input for the local outlier factor (LOF) algorithm to create a model for each subject, with the aim of learning from it and rejecting instances not related to the subject in the model. In search of a minimal subset of EEG channels, we used a channel-selection method based on the non-dominated sorting genetic algorithm (NSGA)-III, designed with the objectives of minimizing the required number EEG channels and increasing the true acceptance rate (TAR) and true rejection rate (TRR). This method was tested on EEG signals from 109 subjects of the public motor movement/imagery dataset (EEGMMIDB) using the resting-state with the eyes-open and the resting-state with the eyes-closed. We were able to obtain a TAR of $$1.000 \pm 0.000$$ 1.000 ± 0.000 and TRR of $$0.998 \pm 0.001$$ 0.998 ± 0.001 using 64 EEG channels. More importantly, with only three channels, we were able to obtain a TAR of up to $$0.993 \pm 0.01$$ 0.993 ± 0.01 and a TRR of up to $$0.941 \pm 0.002$$ 0.941 ± 0.002 for the Pareto-front, using NSGA-III and DWT-based features in the resting-state with the eyes-open. In the resting-state with the eyes-closed, the TAR was $$0.997 \pm 0.02$$ 0.997 ± 0.02 and the TRR $$0.950 \pm 0.05,$$ 0.950 ± 0.05 , also using DWT-based features from three channels. These results show that our approach makes it possible to create a model for each subject using EEG signals from a reduced number of channels and reject most instances of the other 108 subjects, who are intruders in the model of the subject under evaluation. Furthermore, the candidates obtained throughout the optimization process of NSGA-III showed that it is possible to obtain TARs and TRRs above 0.900 using LOF and DWT- or EMD-based features with only one to three EEG channels, opening the way to testing this approach on bigger datasets to develop a more realistic and usable EEG-based biometric system.