Fraud Detection in Banking using Key Agreement and Face Authentication

In today's digital age, the convenience of online financial transactions is accompanied by rising cybersecurity risks. The shift to digital platforms for conducting daily financial activities has significantly exposed customers to cyber threats. These threats often result from unauthorized access, phishing attacks, and fraud attempts by cybercriminals. Ensuring the security of these transactions is crucial to maintain users' trust and to protect sensitive financial data. This study proposes a comprehensive security framework to strengthen the safety of online financial transactions, using a combination of multi-factor authentication (MFA) and a machine learning-based fraud detection system, termed "E-cyberboost." The framework operates on two primary layers, each designed to address a different aspect of security: identity verification and real-time fraud detection. The first layer involves multi-factor authentication, where users must verify their identity. This layer typically combines something the user knows (like a password) with something they have (such as a one-time password or OTP sent to their mobile device) or something they are (such as fingerprint or facial recognition). MFA significantly reduces unauthorized access by ensuring that even if one authentication factor is compromised, additional verification is required to gain access. The second layer introduces an advanced machine-learning component. Named "E-cyberboost," this machine learning model actively monitors ongoing transactions, detecting patterns that may indicate fraudulent activity. If the system identifies any unusual or suspicious transaction behaviour, it immediately triggers a security response, such as additional user verification or temporarily blocking the transaction. This proactive approach allows the system to address potential fraud in real time, adapting to evolving threats by learning from previous incidents and continuously improving its detection accuracy. By combining the preventive measures of MFA with the responsive capabilities of machine learning, this framework provides robust security against both common and sophisticated cyber threats. The study discusses the methodology, implementation, and effectiveness of this layered approach in reducing the risk of unauthorized transactions and protecting users' financial information. The proposed framework aims to enhance trust in online financial platforms and ensure a safer digital transaction experience for users. Index Terms—Multifactor authentication, machine learning, fraud detection, E-cyberboost

To enhance the security and authentication processes, this extension integrates Elliptic Curve Cryptography (ECC) for image encryption and Local Binary Pattern Histogram (LBPH) for live face authentication during registration and login. ECC, known for its high security with smaller key sizes, is incorporated to securely encrypt images, addressing the growing need for efficient and lightweight image encryption in resource-constrained environments. By employing ECC, the encryption algorithm ensures that visual data transmitted and stored remains confidential, with reduced computational overhead compared to traditional encryption methods like AES. Additionally, LBPH is employed for real-time face authentication, allowing users to authenticate themselves via facial recognition during both registration and login. This biometric system leverages LBPH’s robustness against variations in lighting and facial expressions, ensuring reliable face recognition in diverse environments. By the integrity of user access control, offering a combining ECC for image encryption and LBPH for face authentication, the extended system strengthens both the confidentiality of transmitted data and seamless and secure authentication process. The enhanced system was evaluated for speed, accuracy, and efficiency, showing significant improvements in real-time security applications.

The use of information technologies has become an integral part of the modern world, and the internet has played a vital role in facilitating access to various services, from ordering food to obtaining government services. However, with the increased reliance on the internet, the issue of information protection has become more pressing than ever. In the wake of the COVID-19 pandemic, remote work and online learning have become the norm, which has further emphasized the need for robust information protection mechanisms. One of the primary ways of ensuring information security is through the use of cryptographic algorithms. These algorithms have been in use for several decades and have provided an effective means of protecting sensitive information. However, with the exponential growth of computing power, the demands on these algorithms have increased, and some of them have become outdated and are no longer secure enough. As a result, there is a pressing need for the development of new cryptographic algorithms that use more complex mathematical principles to provide a higher level of security. One such class of algorithms that has gained popularity in recent years is elliptic curve cryptography. These algorithms use the principles of elliptic curves to provide greater security while using relatively fewer system resources. Elliptic curve cryptography has several advantages over other cryptographic algorithms. Firstly, it provides a higher level of security with shorter key lengths, making it ideal for use in devices with limited processing power and memory. Additionally, it offers greater resistance to attacks by quantum computers, which are expected to become more prevalent in the future. Moreover, elliptic curve cryptography has found applications in several areas, including secure communication protocols, digital signatures, and key exchange algorithms. These algorithms are part of the TLS protocol, which is used by the HTTPS protocol to ensure secure transmission of information over the Internet. The continuous development and integration of cryptographic algorithms into various information systems will be crucial for protecting against cyber threats in the future.

In recent times, Debiao et al. proposed an ID-based user authentication with key agreement protocol on elliptic curve cryptography (ECC). Conversely, our analysis shows that their scheme contains various security limitations such as many logged-in users' problem, privileged insider attack, impersonation attack and so on. In this paper, we propose an anonymous ID-based remote mutual authentication with key agreement protocol on ECC using smart cards. The proposed scheme affords important security features such as identity protection, anonymity and secure session-key while remain the merits of Debiao's scheme. Our scheme guards the original identity of user by making it anonymous using pseudo random number generators, in order to free it from re-registration problem and smart-card-stolen attacks. The proposed scheme is also a two-factor authentication scheme, and makes the authentication procedure more secure by asking the user to provide two-means of identification. It also provides identity updating feature as depicted in further subsections.

In this work, a facial authentication system has been included examined, deployed and evaluated for students who study in distance higher education. Until today, there was hardly any technology that verified whether the students are themselves the ones in the realization of their activities within a LMS (Learning Management System) as Moodle. Currently, there are different technologies that identify people. One of them is biometrical facial authentication, which permits the authentication and verification of users based on facial features. Thanks to the new technologies’ contribution within the educational community, there is a possibility to verify that there are not frauds and to avoid the identity theft while the students do their activities in the platform; in order to demonstrate that e-learning is as accepted and recognized as any other kind of education. The main objective is to check the functionality degree of facial authentication of the Smowl tool in Moodle platform, to identify the attention degree and opinions in the students’ understanding, as well as to determine which kinds of activities have more acceptance before and after using the tool. In this work, 100 Master students were audited for the survey and the conclusions reached indicate a high acceptance of facial authentication as a technique to improve distance education, and that this perception of students improves even more after experiencing facial authentication when learning through Moodle rooms.KeywordsFacial RecognitionDistance EducationLearn Management SystemIdentity TheftEducational Data MiningThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Cryptography is a technique used today hiding any confidential information from the attack of an intruder. Today data communication mainly depends upon digital data communication, where prior requirement is data security, so that data should reach to the intended user. The protection of multimedia data, sensitive information like credit cards, banking transactions and social security numbers is becoming very important. The protection of these confidential data from unauthorized access can be done with many encryption techniques. So for providing data security many cryptography techniques are employed, such as symmetric and asymmetric techniques. In this review paper different asymmetric cryptography techniques, such as RSA (Rivest Shamir and Adleman), Diffie-Hellman, DSA (Digital Signature Algorithm), ECC (Elliptic curve cryptography) are analyzed. Also in this paper, a survey on existing work which uses different techniques for image encryption is done and a general introduction about cryptography is also given. This study extends the performance parameters used in encryption processes and analyzing on their security issues.

Quantum Machine Learning (QML) is emerging as a transformative technology in cybersecurity, particularly in anomaly detection for cyber security audits. Traditional machine learning models are effective but face scalability and efficiency limitations as cyber threats grow more sophisticated. QML, leveraging quantum computing’s ability to process and analyze large datasets in parallel, offers potential breakthroughs in identifying anomalous patterns that could signify cyber threats such as data breaches, insider threats, or unauthorized access. Content Analysis Research Methodology used in this research work. This paper explores the integration of QML into anomaly detection systems for cyber security audits, where detecting deviations from normal behavior is crucial. Quantum algorithms, particularly those based on Quantum Support Vector Machines (QSVM), Quantum Neural Networks (QNN), and Quantum Principal Component Analysis (QPCA) can enhance the detection of subtle anomalies that classical algorithms may overlook due to noise or the complex, high-dimensional nature of cyber data. The inherent properties of quantum computing, such as superposition and entanglement, allow for more efficient feature selection and optimization, potentially leading to faster and more accurate anomaly detection. The impact of implementing QML in cyber security audits is profound. First, it enhances detection capabilities by identifying anomalies with greater precision, reducing false positives, and improving response times to cyber incidents. Second, quantum algorithms’ ability to manage exponentially large datasets makes them ideal for environments with extensive data logs, such as enterprise networks and cloud infrastructures. Third, as cyber threats become increasingly adaptive and stealthy, QML offers a dynamic solution that evolves alongside these threats by continuously learning from new patterns of attack. However, practical challenges remain, including the need for quantum hardware advancements, the development of hybrid quantum-classical models, and ensuring the interpretability of quantum models in audit scenarios. Despite these challenges, early research and experimental implementations demonstrate the potential of QML to revolutionize anomaly detection in cybersecurity audits. This paper concludes that while QML is still in its early stages, its application to anomaly detection holds promise for significantly enhancing the effectiveness of cyber security audits. The impact of this technology, when fully realized, could redefine how organizations protect their networks and data from ever-evolving cyber threats, making QML a critical area for future research and development in cybersecurity.

The Internet of Things (IoT) is one of the emerging fields of technology, which allows the appliance of Smart Home to connect via the internet for communicating and sharing data between them. However, ensuring the security of IoT devices is challenging for their processing and power consumption. IoT devices are resource-constrained, and traditional encryption algorithms are computationally expensive for these small-size and low-powered devices. In this research, an efficient and effective lightweight cryptographic scheme is presented for securing communication between IoT Home devices based on Elliptic Curve Cryptography (ECC). It ensures complete protection against security risks such as authentication, confidentiality, integrity, and key agreement and protects against common cyber attacks. We have conducted an experimental evaluation to demonstrate that the proposed method can outperform the state-of-the-art cryptographic techniques, which significantly improves the security and reduces the storage and communication cost.KeywordsIoTSmart HomeConfidentialityECCSecurityLightweight cryptographyAuthenticationIntegrity

Using an extended technology acceptance model to investigate facial authentication

The integration of cloud computing into healthcare systems has revolutionized data management by enabling seamless data sharing and enhanced accessibility, thereby improving patient outcomes and operational efficiency. However, the sensitive nature of healthcare data necessitates robust privacy-preserving mechanisms to protect patient confidentiality. This paper explores state-of-the-art techniques such as cryptography, data anonymization, and secure access control, which are tailored to address privacy challenges in healthcare environments. It also critically evaluates existing frameworks, identifying limitations, and proposes a scalable architecture designed to ensure secure and efficient data sharing. Healthcare systems increasingly rely on cloud environments for storing and sharing sensitive patient data. However, these systems face significant challenges, including the risk of data breaches, unauthorized access, and the need for strict compliance with regulatory frameworks such as HIPAA and GDPR. To address these challenges, this paper presents a privacy-preserving framework that leverages advanced cryptographic methods, blockchain for secure transaction logging, and machine learning-based anomaly detection to safeguard patient data from collection to sharing. Experimental results demonstrate that the proposed system outperforms traditional methods in terms of privacy, efficiency, and scalability, showcasing its potential for real-world healthcare applications. The integration of cloud computing into healthcare systems has significantly transformed how patient data is managed and shared across multiple stakeholders, including healthcare providers, insurers, and researchers. This digital transformation has enabled real-time access to medical records, improving clinical decision-making and patient outcomes. However, the sensitive nature of healthcare data demands the adoption of advanced privacy-preserving techniques to mitigate risks such as unauthorized access, data breaches, and regulatory non-compliance. In this paper, we explore state-of-the-art technologies like cryptographic techniques (homomorphic encryption, differential privacy), data anonymization (k-anonymity, l-diversity), and access control mechanisms (RBAC, ABAC) that have been proposed to enhance data privacy and security. Additionally, we delve into the use of blockchain technology for secure, transparent transaction logging and machine learning-based anomaly detection systems to safeguard against unauthorized access. Through experimental analysis and performance testing, we demonstrate that the proposed privacy-preserving framework outperforms conventional methods in terms of scalability, efficiency, and data protection, thereby offering a promising solution for real-world healthcare applications.

Chest X-ray images are an essential diagnostic tool in the medical field, but transmitting and storing these images can raise security concerns. Privacy-preserving techniques in healthcare databases protect sensitive patient information from unauthorized access, breaches, and misuse. Using a combination of elliptic curve cryptography (ECC) for encryption, a convolutional neural network (CNN) for feature extraction, and a long short-term memory (LSTM) network for classification, the aim of this research is to develop a safe and effective classification system for chest X-ray images. The research integrates hyperparameter optimization, essential for LSTM, as it dramatically impacts the model's performance. Additionally, the research employs an oppositionbased learning strategy, the traditional wild horse optimization (WHO) approach and an enhanced WHO (EWHO), to enhance the classification performance. The research objectives are to diagnose and classify chest X-ray images into four types: normal lung, COVID-19-infected lung, lung opacity, and viral pneumonia. The proposed opposition-based learning model improves the accuracy, efficiency, and consistency of diagnostic and treatment decisions. The proposed EWHO-configured LSTM approach attains 97.4% accuracy for an efficient classification system for chest X-ray images. The research highlights the importance of privacy-preserving techniques in healthcare databases to protect sensitive patient information from unauthorized access, breaches, and misuse.

With hardware support and software optimization, Public Key Cryptography (PKC) has been announced feasible on micro sensors recently [1][2][22][23]. A number of experiments proved that the Elliptic Curve Cryptography (ECC) is more suitable for resource constraint motes compared with RSA [2][3]. But even ECC based protocols still cost too much energy. In this paper, we propose C4W, an identity-based public key infrastructure specially designed for wireless sensor networks (WSNs), in which all nodes can generate others' ECC public keys directly from their identities. Without certificates, no energy will be consumed for certificates communication and verification, which makes C4W especially energy efficient. C4W uses a protocol without certificates to realize mutual authentication and key agreement. Compared with a simplified SSL (SSSL) protocol using an abbreviated certificate, C4W consumes lower than 35% energy, and the communication consumption of C4W is only 28.5% of that consumed by SSSL. Furthermore, the energy analysis of C4W illuminates that the expensive public key computational cost is almost neglectable compared with the heavy communication consumption in a large-scale WSNs, which gives the asymmetric key management in WSNs a bright future.

Mobile banking applications have revolutionized financial services by offering consumers quick and convenient access to their accounts. However, this convenience comes with significant security risks, including phishing attacks, data breaches, and unauthorized access. This study investigates these security challenges and explores mitigation strategies through the application of various algorithms. The focus is on two key areas: intrusion detection and data encryption. We assess the effectiveness of Convolutional Neural Networks (CNNs) and Support Vector Machines (SVMs) in detecting malicious activities. Our results indicate that CNNs outperform SVMs by achieving a 15% higher accuracy in identifying complex attacks, owing to their enhanced ability to interpret intricate patterns. In the domain of data encryption, we compare Elliptic Curve Cryptography (ECC) with the Advanced Encryption Standard (AES). While AES excels in terms of speed and robustness, ECC offers a more efficient solution with comparable security levels. Based on our findings, we recommend the adoption of hybrid encryption schemes and enhanced anomaly detection systems as potential mitigation strategies. The study underscores the importance of integrating multiple security approaches to protect mobile banking applications from evolving threats.

The Evolved Packet System Authentication and Key Agreement (EPS-AKA) scheme provides security in Long Term Evaluation (LTE) network. However, EPS-AKA is vulnerable to various security issues such as Denial of Service Attack (DoS), replay attack and lack of privacy protection. Moreover, it does not provide forward secrecy. To remove these security flaws, this paper aims for an Enhanced Secure Authentication and Key Agreement scheme (ES-EPS-AKA) for LTE networks based on Elliptic Curve Cryptography (ECC) technique. The proposed scheme also employs different encrypted functions and HMAC. Security analysis of the proposed ES-EPS-AKA has been analyzed and compared with EPS-AKA and other related existing protocols with respect to various security attributes. Furthermore, the performance of the ES-EPS-AKA has been evaluated and compared with existing protocols in terms of communication, storage and computational overhead. Results show that the proposed ES-EPS-AKA outperforms the existing schemes with respect to above said parameters.

The modern realities of the digital transformation of society make it critically important to ensure cybersecurity as a key element of the state's national security. Ukraine, in conditions of permanent confrontation with hybrid threats and active digitalization of state and private structures, is faced with the need to form an effective system of legal regulation in the field of cyber defense. This study is aimed at a comprehensive analysis of the current state of legislative support for cybersecurity in Ukraine and identifying priority areas for its improvement. The relevance of the study is due to the rapid development of information and communication technologies, the increase in the number and sophistication of cyber attacks, as well as the need to adapt national legislation to European cybersecurity standards. The analysis of modern cyber threats demonstrates their cross-border nature, a high degree of coordination and the use of advanced technologies to cause damage to critical infrastructure, state institutions and the private sector. Special attention is paid to the study of the specifics of cyberterrorism, information operations and attacks on critical infrastructure in the context of modern geopolitical challenges. The methodological basis of the study includes a comprehensive approach to the analysis of regulatory legal acts, international agreements and the practice of their implementation into national legislation. The methods of comparative law, system analysis and empirical methods of research into the practice of law enforcement in the field of cybersecurity were used. Special attention is paid to the study of the experience of the European Union countries in the field of legal regulation of cyber protection and the possibilities of its adaptation to Ukrainian realities. The results of the study indicate the presence of significant gaps in the current legislation of Ukraine on the regulation of cybersecurity, including insufficient coordination between various state bodies, the absence of clear mechanisms for responding to cyber incidents and the inadequacy of sanctions for cyber offenses. Problems in the field of personal data protection, ensuring the security of critical information infrastructure and international cooperation in the field of countering cybercrime were identified. It was established that the existing regulatory framework does not fully meet the modern challenges of the digital age and requires comprehensive reform. The practical significance of the study lies in the development of specific recommendations for improving the legislative regulation of cybersecurity in Ukraine, including proposals for creating an integrated system of cyber protection, improving mechanisms for interagency interaction and increasing the effectiveness of countering cyber threats. Models of legal regulation adapted to the specifics of the Ukrainian legal system and compatible with international cybersecurity standards are proposed. Prospects for further research are related to the need for constant monitoring of the evolution of cyber threats and the appropriate adaptation of the legislative framework, the development of mechanisms for public-private partnership in the field of cybersecurity, as well as an in-depth study of the impact of artificial intelligence and new technologies on the system of legal regulation of cyber defense. An important direction is also the study of the role of international cooperation in the formation of the global cybersecurity architecture and the place of Ukraine in this system.