![ Md. Mahmudul Hasan
Prof. Dr. Engr. Mohiuddin Ahmad
Department of Electrical and Electronic Engineering, KUET, Khulna-9203, Bangladesh.
Contribution
Working Principle Classification & Authentication
Feature Extraction
Development of a EEG-Based Biometric Authentication &
Security System
Motivation and Objectives
Development of a biometric
authentication system based on
electroencephalogram(EEG).
Comparison of EEG based
authentication with latest
technologies such as:
Radio frequency Identification(RFID)
Near Field Communication(NFC).
Face detection using Image
Processing.
To fill the gap between conventional
authentication system and biometric
authentication.
Step 1: Signal Acquisition
Step 2: Row signal pre-processing
a) Filtering
b) Artifact Removal
Step 3: Feature Extraction
Time-domain feature extraction (mean, median,
and variance)
Frequency-domain feature extraction (WT based)
and
Time-frequency domain feature extraction(WPD
i.e. multi-level filters to analyze the time-
frequency information)
Step 4: Classification
SVM(support vector machine)
ANN(Artificial neural Network)
Step 5: Biometric Authentication
Fig 4. Plotting channel spectra
.
Discussion and Conclusion
References
System Overview
EEG & Human Authentication
Flow Chart
Fig. 1. Flow chart of the EEG based security system.
EEG:
Short form of “Electroencephalography”.
EEG is nothing but human
“brainwave signal”.
EEG brainwave pattern is unique from
one person to another like fingerprint.
EEG signals are the signature of neural activities.
It is almost impossible to steal.
EEG is very difficult to mimic.
EEG is referred as “Language of Mind”.
1) Something you know:
Textual password
You have a secret that you know
2) Something you have:
A smart card
Barcode
3) Something you are:
A fingerprint
Iris
EEG signal
EEG is a unique biometric
identifier.
EEG is Just like Fingerprint.
EEG Can not be easily
duplicated.
Adjustments of the EEG head set
is very much important.
FIR filter is used to remove linear
trends.
Individuals with bushy hair don’t
have any problems.
according to the EEG Headset.
EEG Headset are costlier.
Effective for strong security
system.
To proof that EEG signal alone is able to
create a unique pattern for each subject.
To use simple feature extraction methods.
To provide simple classification methods
to provide strong evidence.
To make a novel algorithm with EEG signal
processing.
To provide unique patterns and identify
people with other human features.
To build a reliable security system.
Authentication system:
(Three fundamental Factors)
Fig. 2. Raw EEG Signals
Fig 3. plotting channel locations
Fig 5. After FIR Filtering (linear trends removed)
Signal Acquisition: EEG Headset
Fig 7. Statistical analysis on the dataset
Fig. 6 : After removing channel artifacts
Channel artifact removal
[1] M. Poulos, M. Rangoussi, V. Chrissikopoulos, A. Evangelou. Parametric
person identification from the EEG using computational geometry. Proceedings
of the IEEE International Conference on Electronics, Circuits, and Systems,
2:1005–1008, 1999.
[2] R. Paranjape, J.Mahovsky, L. Benedicenti, Z. Koles. The
electroencephalogram as a biometrics. Proceedings of the Canadian Conference
on Electrical and Computer Engineering, 2:1363–1366, 2001.
[3] R. Palaniappan, D. Mandic. Biometrics from brain electrical activity: A machine
learning approach. IEEE Transactions on Pattern Analysis and Machine
Intelligence, 29(4):738–742, April 2007.
[4] S. Marcel, J. R. Mill´an. Person authentication using brainwaves (EEG) and
maximum a posteriori model adaptation. IEEE Transactions on Pattern Analysis
and Machine Intelligence, 29(4):743–748, April 2007.
[5] Sebastien Marcel a Jose del R. Millan Person authentication using
brainwaves (EEG) and Maximum a posteriori model adaptation, IEEE
Transactions on Pattern Analysis and Machine Intelligence Special Issue on
Biometrics 2007
[6] Xu Huang, Salahiddin Altahat, Dat Tran, Dharmendra Sharma, Human
Identification withElectroencephalogram (EEG) Signal Processing, 2012
International Symposium on Communications and Information Technologies
(ISCIT)](https://image.slidesharecdn.com/eegbasedsecurity3-160129090742/85/Development-of-a-EEG-Based-Biometric-Authentication-Security-System-1-320.jpg)
Development of a EEG-Based Biometric Authentication & Security System
- 1. Md. Mahmudul Hasan Prof. Dr. Engr. Mohiuddin Ahmad Department of Electrical and Electronic Engineering, KUET, Khulna-9203, Bangladesh. Contribution Working Principle Classification & Authentication Feature Extraction Development of a EEG-Based Biometric Authentication & Security System Motivation and Objectives Development of a biometric authentication system based on electroencephalogram(EEG). Comparison of EEG based authentication with latest technologies such as: Radio frequency Identification(RFID) Near Field Communication(NFC). Face detection using Image Processing. To fill the gap between conventional authentication system and biometric authentication. Step 1: Signal Acquisition Step 2: Row signal pre-processing a) Filtering b) Artifact Removal Step 3: Feature Extraction Time-domain feature extraction (mean, median, and variance) Frequency-domain feature extraction (WT based) and Time-frequency domain feature extraction(WPD i.e. multi-level filters to analyze the time- frequency information) Step 4: Classification SVM(support vector machine) ANN(Artificial neural Network) Step 5: Biometric Authentication Fig 4. Plotting channel spectra . Discussion and Conclusion References System Overview EEG & Human Authentication Flow Chart Fig. 1. Flow chart of the EEG based security system. EEG: Short form of “Electroencephalography”. EEG is nothing but human “brainwave signal”. EEG brainwave pattern is unique from one person to another like fingerprint. EEG signals are the signature of neural activities. It is almost impossible to steal. EEG is very difficult to mimic. EEG is referred as “Language of Mind”. 1) Something you know: Textual password You have a secret that you know 2) Something you have: A smart card Barcode 3) Something you are: A fingerprint Iris EEG signal EEG is a unique biometric identifier. EEG is Just like Fingerprint. EEG Can not be easily duplicated. Adjustments of the EEG head set is very much important. FIR filter is used to remove linear trends. Individuals with bushy hair don’t have any problems. according to the EEG Headset. EEG Headset are costlier. Effective for strong security system. To proof that EEG signal alone is able to create a unique pattern for each subject. To use simple feature extraction methods. To provide simple classification methods to provide strong evidence. To make a novel algorithm with EEG signal processing. To provide unique patterns and identify people with other human features. To build a reliable security system. Authentication system: (Three fundamental Factors) Fig. 2. Raw EEG Signals Fig 3. plotting channel locations Fig 5. After FIR Filtering (linear trends removed) Signal Acquisition: EEG Headset Fig 7. Statistical analysis on the dataset Fig. 6 : After removing channel artifacts Channel artifact removal [1] M. Poulos, M. Rangoussi, V. Chrissikopoulos, A. Evangelou. Parametric person identification from the EEG using computational geometry. Proceedings of the IEEE International Conference on Electronics, Circuits, and Systems, 2:1005–1008, 1999. [2] R. Paranjape, J.Mahovsky, L. Benedicenti, Z. Koles. The electroencephalogram as a biometrics. Proceedings of the Canadian Conference on Electrical and Computer Engineering, 2:1363–1366, 2001. [3] R. Palaniappan, D. Mandic. Biometrics from brain electrical activity: A machine learning approach. IEEE Transactions on Pattern Analysis and Machine Intelligence, 29(4):738–742, April 2007. [4] S. Marcel, J. R. Mill´an. Person authentication using brainwaves (EEG) and maximum a posteriori model adaptation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 29(4):743–748, April 2007. [5] Sebastien Marcel a Jose del R. Millan Person authentication using brainwaves (EEG) and Maximum a posteriori model adaptation, IEEE Transactions on Pattern Analysis and Machine Intelligence Special Issue on Biometrics 2007 [6] Xu Huang, Salahiddin Altahat, Dat Tran, Dharmendra Sharma, Human Identification withElectroencephalogram (EEG) Signal Processing, 2012 International Symposium on Communications and Information Technologies (ISCIT)