How Machine Learning Enhances Biometric Authentication

By: KV Sai Mounish, Department of computer science and technology, Student of computer science and technology, Madanapalle Institute Of Technology and Science, 517325, Angallu , Andhra Pradesh.

ABSTRACT –

So electroencephalogram is a device which is used to measure overall electrical activity of brain which will be used in biometric authentication and it will protect against all type of malware attacks due uniqueness of the brain signal everyone have. Recently it has shown that combining it with machine learning has improved biometric authentication process. So we will be giving detailed comparison of machine learning algorithms involved with it. Also explore if any further improvement should be there in future. And those comparisons includes characteristics such as features of data set, methods used, etc. So overall this paper explores in detail explanation of how machine learning promotes biometric authentication and techniques involved to improve accuracy of it in future.

KEYWORDS –

Electroencephalogram, Biometric authentication, Electrical activity, Machine Learning.

INTRODUCTION –

With the increasing importance of secured authentication and a method which is also convenient, Biometric authentication has become an essential factor in recent security systems. However, among different biometric methods EEG-based classifications are very comfortable for personal authentication because even without any secrets identity of subject is available due to brain signals with unique features like maximum uniqueness and complexity. Since the EEG- a record of overall electrical activity in brains, is powerful to differ individuals.

For this reason, machine learning algorithms have proven to be highly effective in examining complex patterns that are underlying in time series[1].

An enhanced comparison of the machine learning algorithms used in EEG-based biometric identification systems is given in this article. We explore the features of the datasets that were utilised, the techniques that were applied, and the performance measures that were attained. We also investigate possible directions for future development with the goal of improving these systems accuracy and even more.

This study attempts to focus on how machine learning progresses the field of biometric identification by a thorough analysis of present methods and potential future developments. It does so by offering insights into the approaches that promote advancement and the potential improvements that may come next.

Background on EEG-Based Biometric Authentication

Overview of EEG Technology is the non-invasive method of measuring the electrical activity of the brain is called electroencephalography (EEG). Electrodes will be applied to our scalp and it will be recording electrical activity, allowing for the capturing of brainwave patterns. EEG signals are divided into various frequency bands, such as delta, theta, alpha, beta, and gamma,[2].

Brain-computer interface (BCI) research, this will br continuously monitoring brain activity with the help of clinical settings. Due to advantage of its higher accuracy it will easilt find out any deviations in brain activity .

Brain signal uniqueness for authentication will be different and character of brain waves is utilized in the biometric authentication process by employing EEG. EEG waves are very difficult to fake compared to other biometric features like fingerprints or facial features.

Individual Variability: Due to changes in the patterns of brain activity the unique features will be differing from and range large upon each individual. By using this every individual will be easily identified by variability.

Dynamic Nature: It is very difficult to manipulate EEG signals since they subject to change over time. Its dynamic quality improves security from efforts at spoofing.

Complexity: Attackers or intruders will be finding difficult to intrude and replicate these signals due to complexity of EEG data.

Combining Machine Learning with EEG

Machine Learning with Biometric Systems is like Biometric systems are highly dependable on machine learning due to its features. Such systems can use ML algorithms to process intricate patterns found in biometric data, which may help make the authentication procedure more secure and reliable. ML is excel in extracting some correlation from these huge dimension, subject-to-subject variable EEG signals, especially when it comes to the authentication based on brainwaves captured with an EEG[3]. By tapping into the unique patterns of brainwaves emitted by each individual, it will further tighten its authentication precision using supervised learning algorithms.
The different character of brain waves is utilized in the biometric authentication process by employing EEG. EEG waves are very difficult to forge compared to other biometric characteristics like fingerprints or facial features because of their inbuilt qualities:

Individual Variability: Due to variations in the structure and function of the brain, EEG patterns range greatly amongst individuals[4]. Every individual can be uniquely identified by this variability.

Dynamic Nature: It is challenging to manipulate EEG signals since they are dynamic and subject to change over time. Its dynamic quality improves security from efforts at spoofing[5].

Complexity: Unauthorised users find it difficult to accurately reproduce the signals due to the great complexity of EEG data[6].

The authentication process involved in EEG based detection including feature extraction, selection shown in Figure 1.

Support Vector Machines (SVM) in EEG-Based Authentication

SVMs are used to find out the best hyperplane lying in the EEG feature space that divides classes, or individuals, because they work well in huge or complex environments.

Neural Networks (NN): Complex temporal patterns in EEG data can be captured by neural networks, especially deep learning models such as convolution neural networks (CNNs)
Decision Trees (DT): Due to their complexity and capacity to manage non-linear relationships in the data, decision trees and ensemble techniques such as Random Forests are utilized.

k-Nearest Neighbours (k-NN): This straightforward and efficient technique classifies EEG signals according to how similar they are to instances that have already been viewed. An supervised learning technique that enhances classification performance is called Random Forest (RF).

So the comparison of different machine learning techniques for EEG based authentication is shown in Table 1.

Table 1: Comparison of machine learning techniques for EEG

CONCLUSION –

So overall there is significant advancement in biometric authentication EEG by integration with machine learning with its properties like robust security, complexity, etc. This application of various algorithms has proved increase in accuracy.

In this article we explored background of EEG its origin and compared various machine learning algorithms including SVM, KNN, Decision tree, Genetic algorithm, Neural networks, etc. Also explored about techniques involved importance of feature extraction, selection.

While EEG based biometric systems offer significant advantage but also has limitations like remove noise, data variability, computation complexity. So the fusion of machine learning with EEG biometric authentication which is highly securable and research continues till now in various ways.

REFERENCES –

1. T. B. Shams, Md. S. Hossain, Md. F. Mahmud, Md. S. Tehjib, Z. Hossain, and Md. I. Pramanik, “EEG-based Biometric Authentication Using Machine Learning: A Comprehensive Survey,” ECTI Trans. Electr. Eng. Electron. Commun., vol. 20, no. 2, pp. 225–241, Jun. 2022, doi: 10.37936/ecti-eec.2022202.246906.
2. J. S. Kumar and P. Bhuvaneswari, “Analysis of Electroencephalography (EEG) Signals and Its Categorization–A Study,” Procedia Eng., vol. 38, pp. 2525–2536, Jan. 2012, doi: 10.1016/j.proeng.2012.06.298.
3. A. K. Singh and S. Krishnan, “Trends in EEG signal feature extraction applications,” Front. Artif. Intell., vol. 5, p. 1072801, Jan. 2022, doi: 10.3389/frai.2022.1072801.
4. O. M. Bazanova and D. Vernon, “Interpreting EEG alpha activity,” Neurosci. Biobehav. Rev., vol. 44, pp. 94–110, Jul. 2014, doi: 10.1016/j.neubiorev.2013.05.007.
5. M. Rahaman, C.-Y. Lin, P. Pappachan, B. B. Gupta, and C.-H. Hsu, “Privacy-Centric AI and IoT Solutions for Smart Rural Farm Monitoring and Control,” Sensors, vol. 24, no. 13, Art. no. 13, Jan. 2024, doi: 10.3390/s24134157.
6. M. Rahaman, F. Tabassum, V. Arya, and R. Bansal, “Secure and sustainable food processing supply chain framework based on Hyperledger Fabric technology,” Cyber Secur. Appl., vol. 2, p. 100045, Jan. 2024, doi: 10.1016/j.csa.2024.100045.
7. Raj, B., Gupta, B. B., Yamaguchi, S., & Gill, S. S. (Eds.). (2023). AI for big data-based engineering applications from security perspectives. CRC Press.
8. Gupta, G. P., Tripathi, R., Gupta, B. B., & Chui, K. T. (Eds.). (2023). Big data analytics in fog-enabled IoT networks: Towards a privacy and security perspective. CRC Press.
9. Chaudhary, P., Gupta, B. B., & Singh, A. K. (2022). XSS Armor: Constructing XSS defensive framework for preserving big data privacy in internet-of-things (IoT) networks. Journal of Circuits, Systems and Computers, 31(13), 2250222.

Cite As

Mounish K.V.S. (2024) How Machine Learning Enhances Biometric Authentication, Insights2Techindo, pp. 1

755100cookie-checkHow Machine Learning Enhances Biometric Authenticationyes