Impact of EEG Frequency Bands and Data Separation on the Performance of Person Verification Employing Neural Networks (original) (raw)
Related papers
On the use of EEG features towards person identification via neural networks
Medical Informatics and The Internet in Medicine, 2001
Person identification based on spectral information extracted from the EEG is addressed in this work -a problem that has not yet been seen in a signal processing framework. Spectral features are extracted non-parametrically from real EEG data recorded from healthy individuals. Neural network classification is applied on these features using a Learning Vector Quantizer in an attempt to experimentally investigate the connection between a person's EEG and genetically -specific information. The proposed method, compared with previously proposed methods, has yielded encouraging correct classification scores in the range of 80% to 100%, (casedependent). These results are in agreement with previous research showing evidence that the EEG carries genetic information. 2 2 addressed to informaticists and computer scientists, aiming to investigate new approaches in the field of EEG analysis. However, it is our belief that results of this study would be of interest to clinicians as well. Potential applications of the proposed person identification method are, for example, information encoding and decoding or access to secure information. EEG recording is non-invasive and medically safe; it therefore constitutes a viable and, under certain conditions, attractive alternative to currently existing forms of person identification based on fingerprints, blood test or retinal scanning. It should be noted, however, that the proposed method is merely indicative rather than deterministic in its classification results, in comparison to DNA identification, e.g., because the biochemical bases of the EEG phenomena are, as yet, completely unknown. Thus, it is not legitimate to claim that EEG identification is equivalent to DNA identification.
Human Electroencephalographic Biometric Person Recognition System
2019
Human head generates various signals according to the situation and activates inside the head as well as outside the head. The frequency of the Head Signal means that brain signal is different as per the level of action taken place by the person; it may be either be imaginary or motor imagery activities. From the brain signals imaginary signals are captured using MindWave Mobile Portable device. Frequency-wise channels are separated and categories as Delta, Theta, Alpha, and Beta. These channels indicated emotions, movement, sensations, vision, etc. Features are extracted of each channel using Power Spectral Density (PSD) function and Deep learning Neural Network. Feature level fusion is used for pattern matching. The Novelty of this work is a single electrode device that is used to capture an Electroencephalography (EEG) imaginary data from the head which is generated by brain functioning. The feature level fusion of channels and Deep learning Neural Network classification of featu...
Sensors
Most studies on EEG-based biometry recognition report results based on signal databases, with a limited number of recorded EEG sessions using the same single EEG recording for both training and testing a proposed model. However, the EEG signal is highly vulnerable to interferences, electrode placement, and temporary conditions, which can lead to overestimated assessments of the considered methods. Our study examined how different numbers of distinct recording sessions used as training sessions would affect EEG-based verification. We analyzed the original data from 29 participants with 20 distinct recorded sessions each, as well as 23 additional impostors with only one session each. We applied raw coefficients of power spectral density estimate, and the coefficients of power spectral density estimate converted to the decibel scale, as the input to a shallow neural network. Our study showed that the variance introduced by multiple recording sessions affects sensitivity. We also showed...
Person identification based on parametric processing of the EEG
1999
Person identification based on parametric spectral analysis of the EEG signal is addressed in this work-a problem that has not yet been seen in a signal-processing framework, to the best of our knowledge. AR parameters are estimated from a signal containing only the alpha, rhythm activity of the EEG. These parameters are used as features in the classification step, which employs a learning vector quantizer network. The proposed method was applied on a set of real EEG recordings made on healthy individuals, in an attempt to experimentally investigate the connection between a person's EEG and genetically-specific information. Correct classification scores at the range of 72% to 84% show the potential of our approach for person classification/identification and are in agreement with previous research showing evidence that the EEG carries genetic information
On the potential of EEG for biometrics: combining power spectral density with a statistical test
International Journal of Biometrics
The objective of this work was to explore the potential of using subject's electroencephalogram (EEG) as a biometric identifier. EEG was collected from eight healthy male participants, while exposing them to the sequence of images displayed on the screen. The averaged, over EEG rhythms, estimates of power spectral density were used as the classification features for the artificial neural network and Euclidean distance-based classifiers. Prior the classification, Kruskal-Wallis test was performed on the power estimates to verify that they were statistically different between different individuals, who were performing identical tasks. Assuming the significance level of 0.075, Kruskal-Wallis analysis indicated that up to 96.42% of such estimates were statistically different between different participants and, therefore, can be used as the classification features for biometric authentication. When using average EEG spectral power as the classification features, the highest classification accuracy of 87.5% was achieved for α 1 EEG rhythm (8-10 Hz), while using the artificial neural network classifier, and for α 2 EEG rhythm (10-14 Hz), while using the Euclidean Distance classifier. The classification performance may be mediated by the type of visual stimulation (i.e., the image the subject perceives) and the statistical test may be instrumental for classification feature selection.
EEG Based User Identification and Verification Using the Energy of Sliced DFT Spectra
2017
Electroencephalogram signals reflect the electrical activity of the brain; EEG signal is the measurement of voltage fluctuations coming from ionic stream within the neurons of the brain. They have been explored in medical researches to diagnose some brain diseases such as Alzheimer's and epilepsy, and have been used in Brian computer interface (BCI) applications. Recently EEG signals are being investigated for identification and verification applications because they show evidence against falsification or replication since the brain activity of people is distinctive. In this paper a promising EEG-based identification and verification system is presented. A feature set based on the energy distribution of Fourier accumulative components is proposed, and some Euclidean distance measures are used for matching. This system was tested on the EEG public CSU dataset which was collected from 7 healthy volunteers. The attained identification results are encouraging with best recognition r...
A Proposed Feature Extraction Method for EEG-based Person Identification
We propose in this paper a feature extraction method to extract brain wave features from electroencephalography (EEG) signal. The proposed feature extraction method is based on an assumption that EEG signal could be considered as stationary if the time window is sufficiently short. With this assumption, EEG signal has some similar properties to speech signal and hence a feature extraction method that is currently used to extract speech features can be applied to extract brain wave features from EEG signal. Mel-frequency cepstral coefficients are features extracted and evaluated in EEG-based person identication. Experimental results show that the proposed method could provide very high recognition rate.
Biometric Identification using Electroencephalography
2020
In this paper, investigate the use of brain activity for person identification. A biometric system is a technological system that uses information about a person. Research on brain signals show that each individual has a unique brain wave pattern. Electroencephalography signals generated by mental tasks are acquired to extract the distinctive brain signature of an individual. Electroencephalography signals during four biometric tasks, namely relax, math, read and spell was acquired from 50 subjects. Features are derived from power spectral density. Classification is performed using Feed forward neural network and Recurrent neural network. The performance of the neural model was evaluated in terms of training, performance and classification accuracies. The results confirmed that the proposed scheme has potential in classifying the EEG signals. RNN is considerably better with an average accuracy of 95% for the spell task and 92% for the read tasks in comparison with a feed forward neu...
Personal identificaion using minimum number of EEG signals
2011
Biometrics such as fingerprints, retinal or iris scanning and face recognition are actively used for identifications. Cognitive biometrics using brain signals have become interesting identification tools because the brain is the most complex biological structure known and its wave signals are very difficult to mimic or steal. In this dissertation, EEG signals are used to identify a person as different persons have different EEG patterns. EEG signals can be measured from different locations. However, many signals can degrade recognition speed and accuracy. A practical technique combining independent component analysis (ICA) for signal cleaning and a supervised neural network for person identification is proposed. From 16 different EEG signal locations, four truly relevant locations of 1,000 data points (F₄, C₄, P₄, O₂), 1,500 data points (F₈, F₃, C₃, P₄), and 3,000 data points (Fp₁, F₄, P₄, O₂) by SOBIRO algorithm were selected. This selection was used to identify a group of 20 perso...
Electroencephalogram (EEG) signals extracted during imagined activities have been studied for use in Brain Computer Interface (BCI) applications. The major hurdle in the EEG based BCI is that the EEG signals are unique to each individual. This complicates a universal BCI design. On the contrary, this disadvantage is the advantage when it comes to using EEG signals from imagined activities for biometric applications. Therefore, in this paper, EEG signals from imagined activities are proposed as a biometric to identify the individuality of persons. The approach is based on the classification of EEG signals recorded when a user performs either one or several mental activities (up to five). As different individuals have different thought processes, this idea would be appropriate for individual identification. To increase the inter-subject differences, EEG data from six electrodes are used instead of one. A total of 108 features (autoregressive coefficients, channel spectral powers, interhemispheric channel spectral power differences and inter-hemispheric channel linear complexity values) are computed from each EEG segment for each mental activity and classified by a linear discriminant classifier using a modified 10 fold cross validation procedure, which gave perfect classification when tested on 500 EEG patterns from five subjects. This initial study has shown the huge potential of the method over existing biometric identification systems as it is impossible to be faked.