False Acceptance Research Articles

MmHSV: In-Air Handwritten Signature Verification via Millimeter-Wave Radar

Electronic signatures are widely used in financial business, telecommuting, and identity authentication. Offline electronic signatures are vulnerable to copy or replay attacks. Contact-based online electronic signatures are limited by indirect contact such as handwriting pads and may threaten the health of users. Consider combining hand shape features and writing process features to form electronic signatures, the article proposes an in-air handwritten signature verification system with millimeter-wave(mmWave) radar, namely mmHSV. First, the biometrics of the handwritten signature process are modeled, and phase-dependent biometrics and behavioral features are extracted from the mmWave radar mixture signal. Secondly, a handwritten feature recognition network based on few-sample learning is presented to fuse multi-dimensional features and determine user legitimacy. Finally, mmHSV is implemented and evaluated with commercial mmWave devices in different scenarios and attack mode conditions. Experimental results show that the mmHSV can achieve accurate, efficient, robust and scalable handwritten signature verification. Area Under Curve (AUC) is 98.96%, False Acceptance Rate (FAR) is 5.1% at the fixed threshold, AUC is 97.79% for untrained users.

An Artificial Intelligence Approach for Verifying Persons by Employing the Deoxyribonucleic Acid (DNA) Nucleotides

Deoxyribonucleic acid (DNA) can be considered as one of the most useful biometrics. It has effectively been used for recognizing persons. However, it seems that there is still a need to propose a new approach for verifying humans, especially after the recent big wars, where too many people lost and die. This approach should have the capability to provide high personal verification performance. In this paper, a personal recognition approach based on artificial intelligence is proposed. This approach is called the artificial DNA algorithm for recognition (ADAR). It utilizes a unique identity for each person acquired from DNA nucleotides, and it can verify individuals efficiently with high performance. The ADAR has been designed and applied to multiple datasets, namely, the DNA classification (DC), sample DNA sequence (SDS), human DNA sequences (HDS), and DNA sequences (DS). For all datasets, a low value of 0% is achieved for each of the false acceptance rate (FAR) and false rejection rate (FRR).

Non-linear optical scattering PUF: enhancing security against modeling attacks for authentication systems.

With the rapid expansion of the Internet of Things (IoT), ensuring the security of personal and group information has become increasingly crucial. However, conventional optical scattering physical unclonable function (OS-PUF) faces challenges due to its linear scattering behavior. In this article, we propose a non-linear OS-PUF (NOS-PUF) that integrates electro-optic materials. By leveraging random refractive index fluctuations generated by the NOS-PUF, we mitigate modeling attacks based on the OS-PUF and bolster the overall security of the authentication process. Moreover, we introduce a novel modeling attack methodology based on scattering invariant modes (SIMs) that poses a significant threat to conventional OS-PUF and NOS-PUF authentication systems. Through extensive simulations, we demonstrate that our NOS-PUF achieves a remarkably lower false accept rate for modeling attacks utilizing SIMs, surpassing the entropy limit imposed by the Gabor filtering algorithm by more than five orders of magnitude. These results highlight the heightened security and increased information entropy offered by the proposed NOS-PUF, making it particularly suitable for applications demanding robust and high-security authentication measures.

PC-based User Continuous Authentication System Using the User's Finger Stroke Characteristics

Biometric technology, which performs continuous authentication based on user behavior, has been developed in various ways depending on the type of device, input device, and sensor. Research on continuous authentication technology in PC-based systems with few sensors installed is based on data from 3physical devices that extract and analyze features from keyboard and mouse input patterns. Among these, previous studies on continuous authentication through keyboard input performed continuous authentication using the key hold delay time that occurs when one key is pressed, the key interval delay time that occurs due to the interaction between fingers, and the key press delay time. did However, the keyboard-based continuous authentication model has limitations in increasing accuracy due to a small number of features. Therefore, in this paper, when a user inputs a sentence using a QWERTY keyboard in a PC system, the function is subdivided by reflecting the characteristics of each finger and used for continuous authentication. The features extracted by reflecting the characteristics of the finger were subdivided into a total of 151 latencies, and SVDD, decision tree, and CNN were used as continuous authentication models. Experimental data was collected through the user's input of randomly displayed sentences, and features were created based on this. User keystroke behavior was used to validate the continuous authentication model. Validation metrics included thresholds for classification accuracy (ACC), ROC curves, false rejection rate (FRR), equal error rate (EER), and false acceptance rate (FAR). As a result of the experiment, it was found that continuous authentication including the user's finger input pattern was superior to the existing method.

Privacy-preserving model for biometric-based authentication and Key Derivation Function

Bio-cryptosystems often save the biometric template for authentication and generally employ randomly generated keys to encrypt and sign data. This method raises privacy protection concerns. Furthermore, for a system secured by the usage of a cryptographic key, losing the key often has disastrous consequences. To overcome the privacy issues, and allow the secure recovery of lost keys, we design a Key Derivation Function to extract a key from biometric data: a new method – based on clustering algorithms – detects consistent and discriminative features from biometric characteristics to create a code. Then, HMAC-SHA256 (as specified by the National Institute of Standards and Technology) generates a standard key from the code. To reproduce the code at future times (that also serves for authentication), the Key Derivation Function stores helper data with the guarantee of privacy. Indeed, with a private face dataset, the probability of generating the code with only the helper data is less than 2−300, and less than 2−246 for a subset of the YouTube Face database. Moreover, on the private database and the tested users from the YouTube Face database, our system has a false acceptance rate of 0%. It corrects up to 40.3% of noise levels on the private database and has good management of the inter-user variability.

An intelligent IoT intrusion detection system using HeInit-WGAN and SSO-BNMCNN based multivariate feature analysis

Devices connected to the Internet of Things (IoT) are growing in popularity across multiple industries. IoT links intelligent homes, buildings, businesses, and communities, and also links computers and mobile devices. Cybersecurity in IoT networks remains challenging, although many intrusion detection services for assault avoidance systems have been developed. Because of enormous growth in computing resources, attack surfaces, communications infrastructure, and attack rates, IoT services encounter substantial security issues. Innovative intrusion-detection systems ensuring data security, availability, and integrity have been developed to mitigate security problems; these systems are recommended for preventing common internal and external assaults and managing IoT environments. To clearly understand any impending attacks, the data are first preprocessed. Missing values are categorized with the concordance connection coefficient, and preprocessing is applied before handling of the missing values. The unevenly distributed and overlapping data are then handled through maximum likelihood estimation-based robust scalar (MLE-RS) scaling followed by random oversampling-based cluster-initialized divisive clustering (ROS-CInitDC). This study conceptualizes a Squirrel search optimization-based Bengio Nesterov momentum convolutional neural network (SSO-BNMCNN)-based approach, which indicates the relationships among information by considering causation. It keeps an optimal processing time and correctness to get data understanding. The selected features are then tested against the categorization of a He initialization-based Wasserstein gradient penalty loss generative adversarial network (HeInit-WGAN). The HeInit-WGAN technique identifies attacks more reliably and accurately than existing methods, reducing false alarms. The proposed strategy had an F-measure of 0.91, and scores of 0.96 for precision, 0.91 for awareness, and 0.98 for explicitness. Moreover, it decreased the false acceptance rate, achieving a 0.02 false positive rate and 0.09 false negative rate.

A Guardbanding Method for Managing False Accept Risk under Process Bias

The test uncertainty ratio (TUR) is often used to ensure false-accept risk is minimal for a given measurement. The commonly used guidance requires either a TUR greater than 4, or appropriate guardbanding, to result in a global false-accept probability of less than 2 %. However, this guidance assumes the distribution of units under test is centered between the tolerance limits and fails to achieve 2 % false accept probability when the product distribution is shifted toward one of the limits. A new guardbanding calculation is proposed that accounts for this potential bias in the product distribution. This guardband method may be applied when no assumptions should be made, and no information is available about the bias of the product distribution and works to ensure a 2 % or lower false accept probability for all TURs (including TURs greater than 4) in the presence of modest product bias.

Toward Multi-User Authentication Using WiFi Signals

User authentication nowadays has become an important support for not only security guarantees but also emerging novel applications. Although WiFi signal-based user authentication has achieved initial success, it works in single-user scenarios while multi-user authentication remains a challenging task. In this paper, we present <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MultiAuth</i> , a multi-user authentication system that can authenticate multiple users with a single pair of commodity WiFi devices. The basic idea is to profile multipath components of WiFi signals, and leverage the multipath components to characterize each user individually for multi-user authentication. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MultiAuth</i> first profiles multipath components of WiFi signals through a proposed MUltipath Time-of-Arrival estimation algorithm (MUTA). Then, after matching corresponding multipath components to each user in complex multi-user scenarios, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MultiAuth</i> constructs individual CSI based on the multipath components to characterize each user individually. An AoA-based approach is exploited to further separate individual CSI constructed by the users with same ToA. To identify users through their activities, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MultiAuth</i> extracts user behavior profiles based on the individual CSI, and leverages a dual-task neural network for robust user authentication. Extensive experiments involving 3 simultaneously present users demonstrate that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">MultiAuth</i> is effective in multi-user authentication with 86.2% average accuracy and 9.5% average false accept rate.

ECG biometric in real-life settings: analysing different physiological conditions with wearable smart textiles shirts

The adoption of biomedical signals such as the electrocardiogram (ECG) for biometric is rising in tandem with the increased attention to wearable devices. However, despite its potential benefits, ECG is rarely implemented as a biometric mechanism in real-life wearable applications. Therefore, this research aims to analyse the ECG signals extracted from wearable Hexoskin Proshirt for biometric authentication in different physiological conditions. A total of 11 subjects participated in this study, where the ECG signals were recorded while standing, sitting, walking, and uncontrolled activity. The raw ECG signal is first pre-processed using noise-removal butterworth filters in the time domain, followed by an efficient QRS segmented feature extraction approach. Finally, around 854 datasets were generated for training and validation, while the remaining 300 were used to test the proposed recognition method with a quadratic support vector machine (QSVM). The results show that the proposed method achieved a reliable accuracy above 98% with false acceptance rate (FAR) of 0.93%, false rejection rate (FRR) of 3.64%, and true positive rate (TPR) above 96% on the in-house datasets. This researchs findings confirm the possibility of using ECG biometrics for authentication purposes in various real-life settings with varying physiological parameters using a smart textile shirt.

Detecting Cyber Attacks In-Vehicle Diagnostics Using an Intelligent Multistage Framework.

The advanced technology of vehicles makes them vulnerable to external exploitation. The current trend of research is to impose security measures to protect vehicles from different aspects. One of the main problems that counter Intrusion Detection Systems (IDSs) is the necessity to have a low false acceptance rate (FA) with high detection accuracy without major changes in the vehicle network infrastructure. Furthermore, the location of IDSs can be controversial due to the limitations and concerns of Electronic Control Units (ECUs). Thus, we propose a novel framework of multistage to detect abnormality in vehicle diagnostic data based on specifications of diagnostics and stacking ensemble for various machine learning models. The proposed framework is verified against the KIA SOUL and Seat Leon 2018 datasets. Our IDS is evaluated against point anomaly attacks and period anomaly attacks that have not been used in its training. The results show the superiority of the framework and its robustness with high accuracy of 99.21%, a low false acceptance rate of 0.003%, and a good detection rate (DR) of 99.63% for Seat Leon 2018, and an accuracy of 99.22%, a low false acceptance rate of 0.005%, and good detection rate of 98.59% for KIA SOUL.

Development of an Improved Multi-filtering Matching Model for Fingerprint Recognition

Over the years research done in the area of fingerprint recognition in which the hybrid matching algorithm is one of the most common techniques, though the hybrid algorithm performed well but still faced with the challenge of false minutiae. This study formulated, simulated, and evaluated a multi-filtering fingerprint matching model to develop a multi-filtering matching model for fingerprint recognition. The method employed a multi-filtering model that was formulated using image pre-processing; minutiae feature extraction, post-processing, and cancellation of false minutiae algorithms in the processed images. The model was simulated using Matlab and fingerprint images from the Fingerprint Verification Competition (FVC) 2002 database. The performance of the model was evaluated using the False Acceptance Rate (FAR), False Rejection Rate (FRR), and Error Equal Rate (EER). The results showed that the false minutiae cancellation algorithm considerably reduced the false minutiae points in the thinned images which resulted in the reduction of false acceptance when two different images were tested, and also reduction in false rejection rate when two same images were tested. The match score was below the threshold value of 50 for false acceptance rate and above the threshold value of 50 for the false rejection rate. The error equal rate EER value of 0.076 was recorded. The study concluded that there was a significant reduction in the false minutiae points present in the thinned images and that a high accuracy of fingerprint matching was achieved when the datasets include poor quality fingerprint images.

Advancements in Speaker Recognition: Exploring Mel Frequency Cepstral Coefficients (MFCC) for Enhanced Performance in Speaker Recognition

Abstract: Speaker recognition, a fundamental capability of software or hardware systems, involves receiving speech signals, identifying the speaker present in the speech signal, and subsequently recognizing the speaker for future interactions. This process emulates the cognitive task performed by the human brain. At its core, speaker recognition begins with speech as the input to the system. Various techniques have been developed for speech recognition, including Mel frequency cepstral coefficients (MFCC), Linear Prediction Coefficients (LPC), Linear Prediction Cepstral coefficients (LPCC), Line Spectral Frequencies (LSF), Discrete Wavelet Transform (DWT), and Perceptual Linear Prediction (PLP). Although LPC and several other techniques have been explored, they are often deemed impractical for real-time applications. In contrast, MFCC stands out as one of the most prominent and widely used techniques for speaker recognition. The utilization of cepstrum allows for the computation of resemblance between two cepstral feature vectors, making it an effective tool in this domain. In comparison to LPC-derived cepstrum features, the use of MFCC features has demonstrated superior performance in metrics such as False Acceptance Rate (FAR) and False Rejection Rate (FRR) for speaker recognition systems. MFCCs leverage the human ear's critical bandwidth fluctuations with respect to frequency. To capture phonetically important characteristics of speech signals, filters are linearly separated at low frequencies and logarithmically separated at high frequencies. This design choice is central to the effectiveness of the MFCC technique. The primary objective of the proposed work is to devise efficient techniques that extract pertinent information related to the speaker, thereby enhancing the overall performance of the speaker recognition system. By optimizing feature extraction methods, this research aims to contribute to the advancement of speaker recognition technology.

Fourier Synchrosqueezing Transform-ICA-EMD Framework Based EOG-Biometric Sustainable and Continuous Authentication via Voluntary Eye Blinking Activities.

In recent years, limited works on EOG (electrooculography)-based biometric authentication systems have been carried out with eye movements or eye blinking activities in the current literature. EOGs have permanent and unique traits that can separate one individual from another. In this work, we have investigated FSST (Fourier Synchrosqueezing Transform)-ICA (Independent Component Analysis)-EMD (Empirical Mode Decomposition) robust framework-based EOG-biometric authentication (one-versus-others verification) performances using ensembled RNN (Recurrent Neural Network) deep models voluntary eye blinkings movements. FSST is implemented to provide accurate and dense temporal-spatial properties of EOGs on the state-of-the-art time-frequency matrix. ICA is a powerful statistical tool to decompose multiple recording electrodes. Finally, EMD is deployed to isolate EOG signals from the EEGs collected from the scalp. As our best knowledge, this is the first research attempt to explore the success of the FSST-ICA-EMD framework on EOG-biometric authentication generated via voluntary eye blinking activities in the limited EOG-related biometric literature. According to the promising results, improved and high recognition accuracies (ACC/Accuracy: ≥99.99% and AUC/Area under the Curve: 0.99) have been achieved in addition to the high TAR (true acceptance rate) scores (≥98%) and low FAR (false acceptance rate) scores (≤3.33%) in seven individuals. On the other hand, authentication and monitoring for online users/students are becoming essential and important tasks due to the increase of the digital world (e-learning, e-banking, or e-government systems) and the COVID-19 pandemic. Especially in order to ensure reliable access, a highly scalable and affordable approach for authenticating the examinee without cheating or monitoring high-data-size video streaming is required in e-learning platforms and online education strategies. Hence, this work may present an approach that offers a sustainable, continuous, and reliable EOG-biometric authentication of digital applications, including e-learning platforms for users/students.

Encrypt with Your Mind: Reliable and Revocable Brain Biometrics via Multidimensional Gaussian Fitted Bit Allocation.

Biometric features, e.g., fingerprints, the iris, and the face, have been widely used to authenticate individuals. However, most biometrics are not cancellable, i.e., once these biometric features are cloned or stolen, they cannot be replaced easily. Unlike traditional biometrics, brain biometrics are extremely difficult to clone or forge due to the natural randomness across different individuals, which makes them an ideal option for identity authentication. Most existing brain biometrics are based on electroencephalogram (EEG), which is usually demonstrated unstable performance due to the low signal-to-noise ratio (SNR). For the first time, we propose the use of intracortical brain signals, which have higher resolution and SNR, to realize the construction of the high-performance brain biometrics. Specifically, we put forward a novel brain-based key generation approach called multidimensional Gaussian fitted bit allocation (MGFBA). The proposed MGFBA method extracts keys from the local field potential of ten rats with high reliability and high entropy. We found that with the proposed MGFBA, the average effective key length of the brain biometrics was 938 bits, while achieving high authentication accuracy of 88.1% at a false acceptance rate of 1.9%, which is significantly improved compared to conventional EEG-based approaches. In addition, the proposed MGFBA-based keys can be conveniently revoked using different motor behaviors with high entropy. Experimental results demonstrate the potential of using intracortical brain signals for reliable authentication and other security applications.

Continuous and Transparent E-Invigilation of E-Assessments

A massive number of users are currently utilizing e-learning. Despite the flexibility provided, the traditional methods of course delivery are still used, including e-examinations. There is significant concern about the potential for cheating. The current solutions in this respect fundamentally fail to offer the required level of security. This paper seeks to develop an e-invigilator that will provide continuous and transparent invigilation of examinees. The study involves a detailed presentation of the proposed architecture and a complete design to be the core of the system, which captures, processes, and monitors students in a controlled and convenient fashion. The proposed framework prototype was developed, presented, and utilized, involving 51 participants to conduct an experiment to explore the viability of the proposed framework. For all 51 participants in this experiment, the false rejection rate was 0 in 2D facial recognition mode, while in 3D facial recognition mode it was 0.04827. Moreover, in order to evaluate the robustness of the approach against targeted misuse, three participants were tasked with a series of nine threat scenarios. The false acceptance rate was 0.038 in the 2D mode and 0 in the 3D mode.

Design of a smart doorbell for a leader’s office with availability status notification and visitor recognition features

Smart doorbells have become a critical component of smart homes and modern offices. However, a smart doorbell, particularly designed for a leader’s office, has not been introduced. In this study, a smart doorbell is developed for a leader’s office. The system includes an application that allows availability status notification on the doorbell module and voice communication with the visitor from inside the office based on a private Wi-Fi network without an Internet connection to prevent the leader from potential privacy and security issues. It also features a live video capture of the visitor with face recognition by implementing a MobileNet model. In training and testing this model, 1,549 free face images of 125 people were augmented to generate training, validation, and testing datasets of 9,185, 2,500, and 5,000 face images, respectively. An additional authentication testing dataset of 1,068 AI-generated face images was also used to evaluate the system’s False Acceptance Rate (FAR). A high confidence level of 0.945 was selected for the developed MobileNet model to obtain zero FAR and high accuracy, recall, and F-score values of 0.960, 0.960, and 0.978, respectively. Therefore, the proposed doorbell could be used for an office leader, showing potential use for biometric authentication.

KeyEncoder: A secure and usable EEG-based cryptographic key generation mechanism

Nowadays, a rapid, easy, and convenient access to our private information is essential to carry out both personal and professional activities. In most cases, this information is sensitive and can be stolen due to its importance and the lack of security protocols. In this study we propose a time–invariant cryptographic key generation mechanism based on electroencephalogram (EEG) signals. We employed Discrete Wavelet Transform and autoencoders to extract the biometric features from the EEG signals. Using these features, we construct a scheme to generate secure seeds that can be used as inputs for secure hash functions and obtain cryptographic keys. The mechanism proposed preserves the privacy of the user, as the cryptographic key is generated for each new EEG signal received, avoiding the need of storing the key, previous EEG signals or any other information. Results show that the proposed mechanism is secure against random attacks, as a 0% of False Acceptance Rate is reported, while generating seeds with an entropy of 0.968 in less than 500 ms.

Passive User Authentication Utilizing Two-Dimensional Features for IIoT Systems

Passive user authentication is critical for the secure operation of Industrial Internet of Things (IIoT) systems. By jointly utilizing both the time-varying characteristics of the user sequential operation actions and spatial variation characteristics of channel state information (CSI) caused by these actions, this paper proposes a novel two-dimensional passive authentication framework for IIoT systems. In particular, we construct the time-varying operation action sequences from the routine work process of a user and apply the Hidden Markov Model to characterize behavioral biometric characteristics of the user, and also employ the eXtreme Gradient Boosting model to depict the spatial variation characteristics of CSI related to the user. By designing two classifiers corresponding these two characteristics and assigning each classifier an appropriate weight, we propose a two-dimensional user authentication framework for continuous and non-intrusive user authentication in IIoT scenarios. Extensive experiments are conducted to illustrate the authentication performance of the proposed authentication framework in terms of false acceptance rate, false rejection rate and equal-error rate. We further investigate the related authentication efficiency issues like the sensitivity to the weights for classifiers, the sensitivity to authentication time and the capability of resisting against impersonation attacks.

Secure UHF RFID Authentication With Smart Devices

Commodity ultra-high-frequency (UHF) RFID authentication systems only provide weak user authentication, as RFID tags can be easily stolen, lost, or cloned by attackers. This paper presents the design and evaluation of SmartRFID, a novel UHF RFID authentication system to promote commodity crypto-less UHF RFID tags for security-sensitive applications. SmartRFID explores extremely popular smart devices and requires a legitimate user to enroll his smart device along with his RFID tag. Besides authenticating the RFID tag as usual, SmartRFID verifies whether the user simultaneously possesses the associated smart device with both feature-based machine learning and deep learning techniques. The user is considered authentic if and only if passing the dual verifications. Comprehensive user experiments on commodity smartwatches and RFID devices confirmed the high security and usability of SmartRFID. In particular, SmartRFID achieves a true acceptance rate of above 97.5% and a false acceptance rate of less than 0.7% based on deep learning. In addition, SmartRFID can achieve an average authentication latency of less than 2.21 s, which is comparable to inputting a PIN on a door keypad or smartphone.

Dual-spectrum network: exploring deep visual feature to attribute mapping for cross-spectral periocular recognition

Recognition of individuals using periocular images has become popular due to its ease of use in situations in which it is difficult to capture full-face regions or high-resolution iris images. Practical surveillance systems recognize individuals from images acquired in various imaging conditions. Recognition of individuals in cross-spectral environments is of practical interest as it can be used to match probe and gallery images that are captured in distinct wavelength domains. The process of matching probe near-infrared images with a gallery of visible images is significantly challenging due to variations in appearance and illumination conditions. In periocular regions, alleviating this wide appearance gap from such local regions is highly challenging. In this work, we propose a dual-spectrum network that considers both high-level deep features and low-level handcrafted attributes to match periocular images in cross-spectral scenarios. A mapping network is designed to establish the relationship between both types of features and generate mapped features that carry discriminative information relevant to both feature domains. The mapped features are passed through a distance layer and subsequent classification block to determine whether they are a genuine pair or an imposter pair. The network is evaluated on a publicly available PolyU dataset as well as two newly created in-house CASIA and IITBBS cross-spectral periocular datasets. The IITBBS dataset consists of visible and near-infrared periocular images from 100 subjects captured in unconstrained imaging conditions with pose and accessory variations. Extensive experiments indicate that the proposed network achieves verification accuracies of 97.96%, 96.48%, and 97.89% on the PolyU, CASIA, and IITBBS datasets, respectively. Performance analysis with existing methods confirms that our network achieves comparable results on the PolyU and CASIA datasets. In addition, the network outperforms existing works on the challenging IITBBS dataset by achieving improvements of 0.43% and 0.16% in terms of verification accuracy and genuine acceptance rate at a 0.1 false acceptance rate.