11th International Conference on Communications, Circuits and Systems

(ICCCAS 2022)
May 13-15, 2022 | Singapore

Invited Speech-20
Use of Spontaneous Blinking for Application in Human
Authentication
Presented by:
Md. Kafiul Islam, PhD, Senior Member IEEE
Associate Professor, Dept. of Electrical & Electronic Engineering
Lab Head, Biomedical Instrumentation and Signal Processing Lab (BISPL)
INDEPENDENT UNIVERSITY, BANGLADESH
Web: https://www.researchgate.net/profile/Md_Kafiul_Islam
Email: kafiul_islam@iub.edu.bd

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 1/15
 Outline
❑ Introduction Reference Article Published in JESTCH, 2020
⮚ Background: Motivation and
Objectives
⮚ EEG, EOG and Eye Blinking
⮚ State-of-the-Art Human
Authentication Systems
❑ Proposed System
⮚ Eye Blink Feature Extraction
⮚ Classification and Statistical Analysis
❑ Performance Evaluation
❑ Comparison with Others
❑ Challenges and Recommendations
❑ Conclusion & Future Work
Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 2/15
 Background: Motivation and Objectives
Objectives:
To develop a robust low-complexity human
authentication system that can not be faked or tampered.

Different Types of Biometric Human Authentication

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 3/15
 EEG and Eye Blinks
What is EEG?
Recording of the brain's spontaneous electrical activity over a period of time
by flat metal discs (electrodes) attached to the scalp. Two major parts of EEG:
1) EEG Rhythms and 2) Transients (such as Epileptic Seizure or Eye Blink artifacts)

EEG Properties Eye Blinks are Considered as

• Amplitude range: 50 uV - 3.5mV
• Broad Frequency range: 0.05 Hz to
100 Hz
Scalp EEG Recording
Artifacts in EEG Recordings
Eye Blinks found in Frontal
EEG Channels

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 4/15
 Proposed System

Simple process flow of the whole method used in this research.

An example of extracted blink signal and

Process flow of the BLINKER tool to extract eye blinks from EEG time series some of its features.

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 5/15
 Data Description and Experiment Setup
Data Collection
• Two Different Setups
1) subject watching a synthetic video clip of 6 min 43 sec
2) subject is asked to like or comment to a series of 150 images shown on a computer screen.
• Emotiv Epoc 14 channel EEG headset with 128 Hz Fs
• Only Fp1 channel’s EEG recording is used EEG Recording:
to extract Eye Blink Signal

• BLINKER- an EEGLAB plugin used

to extrac the Blink Properties

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 6/15
 Analysis of Features

Variable Importance
The Q-Q plot of the amplitude of positive peak of six randomly selected subjects

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 7/15
 Features: Statistical Analysis (Cont.…)

The Kolmogorov-Smirnov test results of 6 randomly chosen subjects for the

duration of positive pulse (top left), amplitude of positive peak of first
The distribution of four most important features
derivative (top right), amplitude of negative peak of first derivative (bottom
for ten randomly selected subjects. left) and the position of positive peak from the onset of positive pulse
(bottom right).

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 8/15
 Results: Classification Accuracy

Boxplot of classification accuracy of GRU with variation in

number of hidden layers and number of blinks in each
sequence.

Boxplot of classification accuracy of six

different classifiers.

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 9/15
 Results: Impact of Blink Features
The impact of features and their values on prediction accuracy

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 10/15
 Results: Accuracy Comparison with Others

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 11/15
 Conclusion
Summary
❑ We investigated the use of spontaneous blink signal for human recognition.
❑ Data of 46 subjects were recorded during a picture visualization task.
❑ After preprocessing the data, 25 features were extracted.
❑ In the next step, the distribution of variables was analyzed to examine whether the generator function is the
same for spontaneous and voluntary blinking.
❑ It was found that features are not distributed identically.
❑ Several models were tested with the purpose of finding the best classifier however, none managed to
distinguish individuals with high accuracy.
❑ Thereby, the possibility of using a sequence model for human authentication was investigated and several
sequences were tested.
❑ Although the best accuracy (98.7%) was achieved by feeding the information carried by a series of 8 blinks
to Gated Recurrent Unit (GRU) Neural Network, our method is capable of using even lesser number of blinks
for accurate human authentication. For future works,

Future Work
The proposed method should be replicated during other tasks such as reading, rest and conversation.

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 12/15
 Selected References
1. Jalilifard, Amir, Dehua Chen, Aunnoy K. Mutasim, M. Raihanul Bashar, Rayhan Sardar Tipu, Ahsan-Ul Kabir Shawon, Nazmus Sakib, M.
Ashraful Amin, and Md Kafiul Islam. "Use of spontaneous blinking for application in human authentication." Engineering Science and
Technology, an International Journal 23, no. 4 (2020): 903-910.
2. Kleifges, Kelly, Nima Bigdely-Shamlo, Scott E. Kerick, and Kay A. Robbins. "BLINKER: Automated extraction of ocular indices from EEG
enabling large-scale analysis." Frontiers in neuroscience 11 (2017): 12.
3. Delorme, Arnaud, and Scott Makeig. "EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics including independent
component analysis." Journal of neuroscience methods 134, no. 1 (2004): 9-21.
4. Emotiv epoc – 14 channel wireless eeg headset.https://www.emotiv.com/ epoc/. Accessed: 2018-08-16..
5. Sherif N. Abbas, M. Abo-Zahhad, Eye blinking EOG signals as biometrics, in: Biometric Security and Privacy, Springer, 2017, pp. 121–
140.
6. Mohammed Abo-Zahhad, Sabah M. Ahmed, Sherif N. Abbas, A novel biometric approach for human identification and verification
using eye blinking signal, IEEE Signal Process. Lett. 22 (7) (2015) 876–880.
7. Mohammed Abo-Zahhad, Sabah M. Ahmed, Sherif N. Abbas, A new multi-level approach to eeg based human authentication using eye
blinking, Pattern Recogn. Lett. 82 (2016) 216–225.
8. Corey Ashby, Amit Bhatia, Francesco Tenore, Jacob Vogelstein, Low-cost electroencephalogram (eeg) based authentication. In Neural
Engineering (NER), in: 2011 5th International IEEE/EMBS Conference on, IEEE, 2011, pp. 442–445.
9. Wu. Qunjian, Ying Zeng, Chi Zhang, Li Tong, Bin Yan, An eeg-based person authentication system with open-set capability combining
eye blinking signals, Sensors 18 (2) (2018) 335.
10. Aunnoy K. Mutasim, Rayhan Sardar Tipu, M. Raihanul Bashar, Md Kafiul Islam, M. Ashraful Amin, Computational intelligence for
pattern recognition in eeg signals, in: Computational Intelligence for Pattern Recognition, Springer, 2018, pp. 291–320.

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 13/15
 Acknowledgment

❑ Professor Renato Assuncção from Brazil for sharing his pearls of wisdom with us during this
research.
❑ My Co-Authors of the article, Especially Mr. Amir Jalilifard from Brazil and his colleague (1st
Author)
❑ My Colleagues Prof. Ashraful Amin, Aunnoy K. Muhtasim and their lab where 2nd EEG dataset
were recorded
❑ Biomedical Instrumentation and Signal Processing Lab (BISPL @ IUB) and its members.
❑ My Dept. of Electrical and Electronic Engineering at IUB.

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 14/15
 Question and Answer

Dr. Md Kafiul Islam Spontaneous Blinking for Human Authentication Page 15/15