publikovani naučni radovi

Naučni rad: "Biometric Authentication Model Based on Transformation of Face Image into a PIN Number Usable During the Covid-19 Pandemic"

Moj naučni rad objavljen u naučnom časopisu za informacionu nauku i tehnologiju ROMJIST koji je indeksiran u "Clarivate Analytics" i to: • Science Citation Index Expanded (također poznat kao SciSearch®), • Journal Citation Reports/Science Edition.
Abstract: The digitization trend is developing throughout the crisis caused by the COVID-19 pandemic. The volume of digital payments is increasing. The most common way of checking the authentication in electronic payment systems is the PIN number that users type into the PINPAD device. Digital payment devices still require the entry of smart cards and the manual entry of a PIN into an ATM or POS device. In order to reduce the possibility of infection due to contamination from multiple touches of the PINPAD device by different people, cardholders use the PINPAD device with gloves. Instead of entering the PIN number on the PINPAD device, biometric authentication is available for the authentication process. On the other hand, the use of gloves and a medical face mask, during the Covid-19 pandemic, limits the biometric scanning of fingerprints and facial images. In this paper, a biometric authentication model is proposed that uses biometric features of the eyes, as these biometric data are available to scanners even in the case when the cardholder uses a protective medical mask on his/her face. The proposed model transforms data on the basis of correlations of characteristic points around the eyes and eye color into a stable PIN. Quantitative presentation of the experimental results confirms that for six different facial expressions of each of the 50 tested persons, the deviation of the authentication PIN from the reference PIN does not exceed 1%. Using the proposed innovative model, the existing infrastructure of payment systems should not be changed.

Keywords: Biometric authentication; Covid-19; electronic payments; facial biometric; PIN number.
Introduction: The authentication model in which the cardholder enters the PIN on the PINPAD device is a widely accepted method of authentication. One of the problems with this method of authentication during the Covid-19 pandemic is the potential source of infection when a cardholder needs to enter the PIN on the PINPAD device or when it needs to be biometrically authenticated by scanning a finger on a biometric scanner. Unhygienic fingers can potentially leave behind surviving bacteria including COVID-19, which is mainly spread by contaminated hands [1]. Research has shown that the COVID-19 virus can last up to 72 hours on plastic and stainless steel [2]. Instead of typing the PIN, where the customer needs to bring a card and touch the machine for electronic data recording, authentication by facial recognition can be developed. In order to avoid the transmission of COVID-19 to card users as a means of payment, it is necessary to develop a new electronic payment model. This paper is a continuation of the author's research published in "A Multimodal Biometric Authentication (MBA) in Card Payment System", 2019 International Conference on Artificial Intelligence: Applications and Innovations (IC-AIAI) [3]. In this paper, a model of transformation of biometric data of a fingerprint and a face image into a multi-digit PIN is presented. A non-invasive biometric method was used, which does not require physical contact with the biometric sensor, and which will be useful in case the cardholder has a covered face with a medical mask as protection against Covid-19 virus infection. The contribution of the research is a software system for PIN authentication that is intended for use during a virus pandemic, but also in other cases when there is a reduced possibility of using traditional PIN authentication methods. The capability of the presented authentication model is in the fact of a simple upgrade of the existing PIN authentication model with a view of users' health care, because they will have a simpler interface of the existing ATM and POS infrastructure with cameras even when wearing a face mask. By applying the presented model of PIN authentication in e-payment systems, the user will not have to enter the authentication PIN, as it will be created from the biometric data of the face image and in the case when wearing a medical mask on his/her face. During the virus pandemic, it is possible for financial institutions to set the PIN authentication model to either the classic method in which the PIN is entered on a PINPAD device, or to the introduced biometric method in which the PIN is calculated from the biometric data of the facial image. Biometric payments are an option that reduces the risk of infection when performing payment transactions [5]. The European Union's Second Payment Services Directive (PSD2) excludes the possibility of applying e-payment authentication models based only on payment card data [4].
New research on e-payment models that will meet the Regulatory Technical Standards on strong customer authentication and secure communication under PSD2 is needed. One of the most innovative technologies, with the greatest growth potential, is based on the use of biometric techniques, with predictions that by 2025 they will be used to authenticate more than $3 trillion in payment transactions, compared to only $404 billion in 2020, and that's an increase of 650% [6]. Payment systems based on facial image recognition have not yet been sufficiently studied and there is no standardization for biometric authentication, but this technology provides the highest security in user authentication [7]. Unlike other biometric attributes, such as fingerprints, their use does not require physical contact with the device. During the Covid-19 pandemic [8] new payment systems have been established to reduce contact between buyers and sellers. The authors in [9] investigated and concluded that there are concerns about the collection of biometric data. The presented method uses a limited set of biometric data, and this has a positive effect on protecting the cardholder's privacy. During the research, the main challenges were the research of existing literature, regulations and standards on biometric cardholder authentication methodologies in electronic payment systems and investigating available biometric SDK software environments solutions. A Python code was developed for a commercial SDK software solution [10] that calculates the IPD distances and the average RGB value of the eye pigment from the face image, as well as the final PIN. The challenge was to research a database of facial images that could be used for research. A professional image database [11] was used, which was created at Binghamton University in New York and is often used by researchers. This paper is conceived as follows: Section 2 presents an overview of previous research. Section 3 presents a conceptual model of the authentication method, and Section 4 describes the algorithm. Section 5 presents the results of experimental research. There is a Future Work given in Section 6, and conclusions are highlighted in Section 7.

[1] G. KAMP, D. TIDT, S. PFAENDER and E. STEINMANN, Persistence of coronaviruses on inanimate surfaces and their inactivation with biocidal agents, Journal of Hospital Infection 104(3), pp. 246–251, 2020.
[2] S. MANIGANDAN, M.-T. Wu, V. K. PONNUSAMY, V. B. RAGHAVENDRA, A. PUGAZHENDHI and K. BRINDHADEVI, A systematic review on recent trends in transmission diagnosis prevention and imaging features of COVID-19, Process Biochemistry 98, pp. 233–240, 2020.
[3] N. BADOVINAC and D. SIMIC, A Multimodal Biometric Authentication (MBA) in Card Payment Systems, Proceedings of 2019 International Conference on Artificial Intelligence: Applications and Innovations, Belgrade, Serbia, pp. 23–26, 2019.
[4] Regulatory Technical Standards on strong customer authentication and secure communication under
PSD2. Accessed: Mar. 10, 2023. [Online]. Available:
[5] F. LIEBANA-CABANILLAS, F. MUNOZ-LEIVA, S. MOLINILLO and E. HIGUERAS-CASTILLO, Do biometric payment systems work during the COVID-19 pandemic? Insights from the Spanish users' viewpoint, Financial Innovation 8(1) paper 22, 2022.
[6] Biometrics to secure over $3 trillion in mobile payments by 2025. Accessed: Mar. 10, 2023. [Online] Available:
[7] A. MODIBBO and Y. ALIYU, Cashless society, financial inclusion and information security in Nigeria: the case for adoption of multifactor biometric authentication, International Journal of Innovative Research in Science Engineering and Technology 4(11), pp. 872–880, 2019.
[8] G. MERTENS, L. GERRITSEN, S. DUIJNDAM, S. SALEMINK and E. ENGELHARD, Fear of the coronavirus (COVID-19): predictors in an online study conducted in March 2020. Journal of Anxiety Disorders 74, paper 102258, 2020.
[9] V. BHARATHAN, Digital dollar project in light of recent congressional hearings, Forbes, Jun. 29, 2020. Accessed: Mar. 10, 2023. [Online]. Available: [10] Commercial Biometric SDK software, Accessed: Mar. 10, 2023. [Online]. Available:
[11] L.-J. YIN, X.-C. CHEN, Y. SUN, T. WORM and M. REALE, A high-resolution 3D dynamic facial expression database, Proceedings of 8
th International Conference on Automatic Face and Gesture Recognition, Amsterdam, Netherlands, pp. 1–6, 2008.
[12] N. CAVUS, Y. B. MOHAMMED, M. BULAMA and M. L. ISAH, Examining user verification schemes, safety and secrecy issues affecting M-banking: Systematic literature review, SAGE Open 13(1), 2023.
[13] J. JAYANTHAN, N. K. PRIYA, S. P. KUMAR and K. SANGEETHA, Facial recognition controlled smart banking, IJRESM 4(3), pp. 185–187, 2021.
[14] A. ZULFIQAR, M. SHAMIM HOSSAIN, M. GHULAM, I. ULLAH, H. ABACHI and A. ALAMRI, Edge-centric multimodal authentication system using encrypted biometric templates, Future Generation Computer Systems 85, pp 76–87, 2018.
[15] K. A. SHAKIL, F. H. ZAREEN, M. ASLAM and S. JABIN, BAMCloud: A cloud based mobile biometric authentication framework, Multimedia Tools and Applications 220, pp. 1–30, 2022.
[33] T. BARBU, Content-based image retrieval framework using multi-scale deeplearning-based feature extraction and graph-based clustering, Proceedings of the Romanian Academy, Series A: Mathematics, Physics, Technical Sciences, Information Science 23(3), 2022.
[34] Android POS machine with fingerprint, card readers, 2SIM and camera. Accessed: Mar. 10, 2023. [Online]. Available:
[35] Y. MERSCH, An ECB digital currency – a flight of fancy, Speech at Consensus Virtual Conference, 2020. Accessed: Mar. 10, 2023. [Online]. Available: 01209cb324.en.html.

Nenad BADOVINAC, Dejan SIMIC (2023): „Biometric Authentication Model Based on Transformation of Face Image into a PIN Number Usable During the Covid-19 Pandemic", Romanian Journal of Information Science and Technology, Volume 26, Number 2, pp. 151–162 | DOI: 10.59277/ROMJIST.2023.2.03